BView

Derived from: BHandler

Declared in: be/interface/View.h

Library: libbe.so

[method summary]

BView objects are the agents for drawing and message handling within windows. Each object sets up and takes responsibility for a particular view, a rectangular area that's associated with at most one window at a time. The object draws within the view rectangle and responds to reports of events elicited by the images drawn.

Classes derived from BView implement the actual functions that draw and handle messages; BView merely provides the framework. For example, a BTextView object draws and edits text in response to the user's activity on the keyboard and mouse. A BButton draws the image of a button on-screen and responds when the button is clicked. BTextView and BButton inherit from the BView class--as do most classes in the Interface Kit.

Views and Windows

For a BView to do its work, you must attach it to a window. The views in a window are arranged in a hierarchy--there can be views within views--with those that are most directly responsible for drawing and message handling located at the terminal branches of the hierarchy and those that contain and organize other views situated closer to its trunk and root. A BView begins life unattached. You can add it to a hierarchy by calling the AddChild() function of the BWindow or of another BView.

Within the hierarchy, a BView object plays two roles:

It's a BHandler for messages delivered to the window thread. BViews implement the functions that respond to the most common system messages--including those that report keyboard and mouse events. They can also be targeted to handle application-defined messages that affect what they view displays.

It's an agent for drawing. Adding a BView to a window gives it an independent graphics environment. A BView draws on the initiative of the BWindow and the Application Server, whenever they determine that the appearance of any part of the view rectangle needs to be "updated." It also draws on its own initiative in response to events.

The relationship of BViews to BWindows and the framework for drawing and responding to the user were discussed in the opening sections of this chapter. The concepts and terminology presented there are assumed in this class description.

BViews can also be called upon to create bitmap images. See the BBitmap class for details.

Locking the Window

If a BView is attached to a window, any operation that affects the view might also affect the window and the BView's shadow counterpart in the Application Server. For this reason, any code that calls a BView function should first lock the window--so that one thread can't modify essential data structures while another thread is using them. A window can be locked by only one thread at a time.

Locking is accomplished through the BLooper Lock() and Unlock() functions inherited by BWindow:

   if ( Window()->Lock() ) {
       . . .
       Window()->Unlock();
   }

Before they do anything else, almost all BView functions check to be sure the caller has the window locked. If the window isn't properly locked, they print warning messages and fail.

Of course, a BView function can require the window to be locked only if the view has a window to lock; the requirement can't be enforced if the BView isn't attached to a window. However, as discussed in "Views and the Server" in the "Drawing" section of this chapter, some BView functions don't work at all unless the view is attached--in which case the window must be locked.

Whenever the system calls a BView function to notify it of something--whenever it calls WindowActivated(), Draw(), MessageReceived() or another hook function--it first locks the window thread. The application doesn't have to explicitly lock the window when responding to an update, an interface message, or some other notification; the window is already locked.

Focus

To facilitate keyboard navigation of views, BView provides integral support for the concept of focus. Only one view in a window can have the focus at any given time, and the view that has the focus is the one whose KeyDown() function is called to process keyboard events.

From the user's point-of-view, the tab key rotates the focus from one view to the next through the navigation group, cycling back to the first view if the tab key is pressed while the last targetable view in the group has the focus. The shift-tab keyboard combination cycles the focus backward. Holding down the control key while cycling groups causes the focus to jump from one navigation group to the next.

When a view has the focus, some sort of indicator should be drawn to inform the user that the view is the focus. Typically, this involves drawing a line under a label in teh view, or possibly drawing a box around the view or some portion of it. The global keyboard_navigation_color() function should be used to obtain the color used to draw the focus indicator.

The view's MakeFocus() function is called to specify whether or not the control has the focus; it's called with an argument of true if the control has the focus, and false if it's not the focus; MakeFocus() calls the previously-focused view's MakeFocus() function to inform that view that it's not the focus anymore. You can augment MakeFocus() is a subclass if you need to take notice when the view becomes the focus (or loses the focus). For example, you may need to draw or erase the keyboard navigation indicator.

Focus is used extensively by controls. For further information, see "Introduction to Controls."

Derived Classes

When it comes time for a BView to draw, its Draw() virtual function is called automatically. When it needs to respond to an event, a virtual function named after the kind of event is called--MouseMoved(), KeyDown(), and so on. Classes derived from BView implement these hook functions to do the particular kind of drawing and message handling characteristic of the derived class.

Some classes derived from BView implement control devices--buttons, dials, selection lists, check boxes, and so on--that translate user actions on the keyboard and mouse into more explicit instructions for the application. In the Interface Kit, BMenu, BListView, BButton, BCheckBox, and BRadioButton are examples of control devices.
Other BViews visually organize the display--for example, a view that draws a border around and arranges other views, or one that splits a window into two or more resizable panels. The BBox, BScrollBar, and BScrollView classes fall into this category.
Some BViews implement highly organized displays the user can manipulate, such as a game board or a scientific simulation.
Perhaps the most important BViews are those that permit the user to create, organize, and edit data. These views display the current selection and are the focus of most user actions. They carry out the main work of an application. BTextView is the only Interface Kit example of such a view.

Almost all the BView classes defined in the Interface Kit fall into the first two of these groups. Control devices and organizational views can serve a variety of different kinds of applications, and therefore can be implemented in a kit that's common to all applications.

However, the BViews that will be central to most applications fall into the last two groups. Of particular importance are the BViews that manage editable data. Unfortunately, these are not views that can be easily implemented in a common kit. Just as most applications devise their own data formats, most applications will need to define their own data-handling views.

Nevertheless, the BView class structures and simplifies the task of developing application-specific objects that draw in windows and interact with the user. It takes care of the lower-level details and manages the view's relationship to the window and other views in the hierarchy. You should make yourself familiar with this class before implementing your own application-specific BViews.

Hook Functions

AllAttached()
Can be implemented to finish initializing the BView after it's attached to a window, where the initialization depends on a descendant view's AttachedToWindow() function having been called.

AllDetached()
Can be implemented to prepare the BView for being detached from a window, where the preparations depend on a descendant view's DetachedFromWindow() function having been called.

AttachedToWindow()
Can be implemented to finish initializing the BView after it becomes part of a window's view hierarchy.

DetachedFromWindow()
Can be implemented to prepare the BView for its impending removal from a window's view hierarchy.

Draw()
Can be implemented to draw the view.

FrameMoved()
Can be implemented to respond to a message notifying the BView that it has moved in its parent's coordinate system.

FrameResized()
Can be implemented to respond to a message informing the BView that its frame rectangle has been resized.

GetPreferredSize()
Can be implemented to calculate the optimal size of the view.

KeyDown()
Can be implemented to respond to a message reporting character input from the keyboard (a key-down event).

KeyUp()
Can be implemented to respond to a message reporting a key-up event.

MakeFocus()
Makes the BView the focus view, or causes it to give up being the focus view; can be augmented to take any action the change in status may require.

MouseDown()
Can be implemented to respond to a message reporting a mouse-down event.

MouseMoved()
Can be implemented to respond to a notification that the cursor has entered the view's visible region, moved within the visible region, or exited from the view.

Pulse()
Can be implemented to do something at regular intervals. This function is called repeatedly when no other messages are pending.

TargetedByScrollView()
Can be implemented to react when the BView becomes the target of a BScrollView.

WindowActivated()
Can be implemented to respond to a notification that the BView's window has become the active window, or has lost that status.

Constructor and Destructor

BView()

      BView(BRect frame, const char *name, uint32 resizingMode, uint32 flags) 
      BView(BMessage *archive)

Sets up a view with the frame rectangle, which is specified in the coordinate system of its eventual parent, and assigns the BView an identifying name, which can be NULL.

When it's created, a BView doesn't belong to a window and has no parent. It's assigned a parent by having another BView adopt it with the AddChild() function. If the other view is in a window, the BView becomes part of that window's view hierarchy. A BView can be made a child of the window's top view by calling BWindow's version of the AddChild() function.

When the BView gains a parent, the values in frame are interpreted in the parent's coordinate system. The sides of the view must be aligned on screen pixels. Therefore, the frame rectangle should not contain coordinates with fractional values. Fractional coordinates will be rounded to the first lower whole number (for example 1.2 will be rounded down to 1.0).

The resizingMode mask determines the behavior of the view when its parent is resized. It should combine one constant for horizontal resizing,

B_FOLLOW_LEFT
B_FOLLOW_RIGHT
B_FOLLOW_LEFT_RIGHT
B_FOLLOW_H_CENTER

with one for vertical resizing:

B_FOLLOW_TOP
B_FOLLOW_BOTTOM
B_FOLLOW_TOP_BOTTOM
B_FOLLOW_V_CENTER

For example, if B_FOLLOW_LEFT is chosen, the margin between the left side of the view and the left side of its parent will remain constant--the view will "follow" the parent's left side. Similarly, if B_FOLLOW_RIGHT is chosen, the view will follow the parent's right side. If B_FOLLOW_H_CENTER is chosen, the view will maintain a constant relationship to the horizontal center of the parent.

If the constants name opposite sides of the view rectangle--left and right, or top and bottom--the view will necessarily be resized in that dimension when the parent is. For example, B_FOLLOW_LEFT_RIGHT means that the margin between the left side of the view and left side of the parent will remain constant, as will the margin between the right side of the view and the right side of the parent. As the parent is resized horizontally, the child will be resized with it. Note that B_FOLLOW_LEFT_RIGHT is not the same as combining B_FOLLOW_LEFT and B_FOLLOW_RIGHT, an illegal move. The resizingMode mask can contain only one horizontal constant and one vertical constant.

If a side is not mentioned in the mask, the distance between that side of the view and the corresponding side of the parent is free to fluctuate. This may mean that the view will move within its parent's coordinate system when the parent is resized. B_FOLLOW_RIGHT plus B_FOLLOW_BOTTOM, for example, would keep a view from being resized, but the view will move to follow the right bottom corner of its parent whenever the parent is resized. B_FOLLOW_LEFT plus B_FOLLOW_TOP prevents a view from being resized and from being moved.

In addition to the constants listed above, there are two other possibilities:

B_FOLLOW_ALL_SIDES
B_FOLLOW_NONE

B_FOLLOW_ALL_SIDES is a shorthand for B_FOLLOW_LEFT_RIGHT and B_FOLLOW_TOP_BOTTOM. It means that the view will be resized in tandem with its parent, both horizontally and vertically.

B_FOLLOW_NONE behaves just like B_FOLLOW_LEFT|B_FOLLOW_TOP; the view maintains the same position in its parent's coordinate system, but not in the screen coordinate system.

Typically, a parent view is resized because the user resizes the window it's in. When the window is resized, the top view is too. Depending on how the resizingMode flag is set for the top view's children and for the descendants of its children, automatic resizing can cascade down the view hierarchy. A view can also be resized programmatically by the ResizeTo() and ResizeBy() functions.

The resizing mode can be changed after construction with the SetResizingMode() function.

The flags mask determines what kinds of notifications the BView will receive. It can be any combination of the following constants:

B_WILL_DRAWIndicates that the BView does some drawing of its own and therefore can't be ignored when the window is updated. If this flag isn't set, the BView won't receive update notifications--its Draw() function won't be called--and it won't be erased to its background view color if the color is other than white.

B_PULSE_NEEDEDIndicates that the BView should receive Pulse() notifications.

B_FRAME_EVENTSIndicates that the BView should receive FrameResized() and FrameMoved() notifications when its frame rectangle changes--typically as a result of the automatic resizing behavior described above. FrameResized() is called when the dimensions of the view change; FrameMoved() is called when the position of its left top corner in its parent's coordinate system changes.

B_FULL_UPDATE_ON_RESIZEIndicates that the entire view should be updated when it's resized. If this flag isn't set, only the portions that resizing adds to the view will be included in the clipping region.

B_NAVIGABLEIndicates that the BView can become the focus view for keyboard actions. This flag makes it possible for the user to navigate to the view and put it in focus by pressing the Tab key. See "Keyboard Navigation" at the beginning of this chapter.

B_NAVIGABLE_JUMPMarks the position of a group of views for keyboard navigation. By pressing Control-Tab, the user can jump from group to group. The focus lands on the first BView in the group that has the B_NAVIGABLE flag set. This may be the same BView that has the B_NAVIGABLE_JUMP marker, or the B_NAVIGABLE_JUMP BView may be the parent of a group of B_NAVIGABLE views.

B_SUBPIXEL_PRECISEInstructs the rendering methods to use subpixel precision when drawing. If you don't set this flag, coordinates are rounded to the nearest unit.

If none of these constants apply, flags can be NULL. The flags can be reset after construction with the SetFlags() function.

See also: SetResizingMode(), SetFlags(), BHandler::SetName()

~BView()

      virtual ~BView()

Frees all memory the BView allocated, and ensures that each of the BView's descendants in the view hierarchy is also destroyed.

It's an error to delete a BView while it remains attached to a window. Call RemoveChild() or RemoveSelf() before using the delete operator.

See also: RemoveChild()

Static Functions

Instantiate()

      static BArchivable *Instantiate(BMessage *archive)

Returns a new BView object, allocated by new and created with the version of the constructor that takes a BMessage archive. However, if the message doesn't contain archived data for a BView, Instantiate() returns NULL.

See also: BArchivable::Instantiate(), instantiate_object()

Member Functions

AddChild(), RemoveChild()

      void AddChild(BView *aView, BView *sibling = NULL) 
      bool RemoveChild(BView *aView)

AddChild() makes aView a child of the BView, provided that aView doesn't already have a parent. The new child is added to the BView's list of children immediately before the named sibling BView. If the sibling is NULL (as it is by default), aView isn't added in front of any other view--in other words, it's added to the end of the list. If the BView is attached to a window, aView and all its descendants become attached to the same window. Each of them is notified of this change through AttachedToWindow() and AllAttached() function calls.

AddChild() fails if aView already belongs to a view hierarchy. A view can live with only one parent at a time. It also fails if sibling is not already a child of the BView.

RemoveChild() severs the link between the BView and aView, so that aView is no longer a child of the BView; aView retains all its own children and descendants, but they become an isolated fragment of a view hierarchy, unattached to a window. Each removed view is notified of this change through DetachedFromWindow() and AllDetached() function calls.

A BView must be removed from a window before it can be destroyed.

If it succeeds in removing aView, RemoveChild() returns true. If it fails, it returns false. It will fail if aView is not, in fact, a current child of the BView.

When a BView object becomes attached to a BWindow, two other connections are automatically established for it:

The view is added to the BWindow's flat list of BHandler objects, making it an eligible target for messages the BWindow dispatches.
The BView's parent view becomes its next handler. Messages that the BView doesn't recognize will be passed to its parent.

Removing a BView from a window's view hierarchy also removes it from the BWindow's flat list of BHandler objects; the BView will no longer be eligible to handle messages dispatched by the BWindow.

See also: BWindow::AddChild(), BLooper::AddHandler(), BHandler::SetNextHandler(), RemoveSelf(), AttachedToWindow(), DetachedFromWindow()

AddLine() see BeginLineArray()

AllAttached() see AttachedToWindow()

AllDetached() see DetachedFromWindow()

AppendToPicture() see BeginPicture()

AttachedToWindow(), AllAttached()

      virtual void AttachedToWindow(void)
      virtual void AllAttached(void)

Implemented by derived classes to complete the initialization of the BView when it's assigned to a window. A BView is assigned to a window when it, or one of its ancestors in the view hierarchy, becomes a child of a view already attached to a window.

AttachedToWindow() is called immediately after the BView is formally made a part of the window's view hierarchy and after it has become known to the Application Server and its graphics parameters are set. The Window() function can identify which BWindow the BView belongs to.

All of the BView's children, if it has any, also become attached to the window and receive their own AttachedToWindow() notifications. Parents receive the notification before their children, but only after all views have become attached to the window and recognized as part of the window's view hierarchy. This function can therefore depend on all ancestor and descendant views being in place.

For example, AttachedToWindow() can be implemented to set a view's background color to the same color as its parent, something that can't be done before the view belongs to a window and knows who its parent is.

   void MyView::AttachedToWindow()
   {
       if ( Parent() )
           SetViewColor(Parent()->ViewColor());
       baseClass::AttachedToWindow();
   }

The AllAttached() notification follows on the heels of AttachedToWindow(), but works its way up the view hierarchy rather than down. When AllAttached() is called for a BView, all its descendants have received both AttachedToWindow() and AllAttached() notifications. Therefore, parent views can depend on any calculations that their children make in either function. For example, a parent can resize itself to fit the size of its children, where their sizes depend on calculations done in AttachedToWindow().

The default (BView) version of both these functions are empty.

See also: AddChild(), Window()

BeginLineArray(), AddLine(), EndLineArray()

      void BeginLineArray(int32 count)
      void AddLine(BPoint start, BPoint end, rgb_color color)
      void EndLineArray(void)

These functions provide a more efficient way of drawing a large number of lines than repeated calls to StrokeLine(). BeginLineArray() signals the beginning of a series of up to count AddLine() calls; EndLineArray() signals the end of the series. Each AddLine() call defines a line from the start point to the end point, associates it with a particular color, and adds it to the array. The lines can each be a different color; they don't have to be contiguous. When EndLineArray() is called, all the lines are drawn--using the then current pen size--in the order that they were added to the array.

These functions don't change any graphics parameters. For example, they don't move the pen or change the current high and low colors. Parameter values that are in effect when EndLineArray() is called are the ones used to draw the lines. The high and low colors are ignored in favor of the color specified for each line.

The count passed to BeginLineArray() is an upper limit on the number of lines that can be drawn. Keeping the count close to accurate and within reasonable bounds helps the efficiency of the line-array mechanism. It's a good idea to keep it less than 256; above that number, memory requirements begin to impinge on performance.

See also: StrokeLine()

BeginPicture(), AppendToPicture(), EndPicture()

      void BeginPicture(BPicture *picture)
      void AppendToPicture(BPicture *picture)
      BPicture *EndPicture(void)

BeginPicture() starts a new "picture recording" session: Subsequent drawing instructions invoked upon the view are recorded in the BPicture argument. AppendToPicture() does the same, but doesn't clear the argument first--it tacks additional instructions on to the end of the BPicture. EndPicture() ends the recording session; it returns the object that was passed to BeginPicture() or AppendToPicture().

While it's recording a picture, the BView doesn't display anything to the screen. To render the drawing, you use the DrawPicture() function.

The picture captures only primitive graphics operations such as DrawString(), FillArc(), and SetFont(). Furthermore, only instructions performed by this view are recorded; the drawing done by the view's children is not recorded.

A BPicture can be recorded only if the BView is attached to a window. The window can be off-screen and the view itself can be hidden or reside outside the current clipping region.

See also: the BPicture class, DrawPicture()

BeginRectTracking(), EndRectTracking()

      void BeginRectTracking(BRect rect, uint32 how = B_TRACK_WHOLE_RECT)
      void EndRectTracking(void)

These functions instruct the Application Server to display a rectangular outline that will track the movement of the cursor. BeginRectTracking() puts the rectangle on-screen and initiates tracking; EndRectTracking() terminates tracking and removes the rectangle. The initial rectangle, rect, is specified in the BView's coordinate system.

This function supports two kinds of tracking, depending on the constant passed as the how argument:

B_TRACK_WHOLE_RECT The whole rectangle moves with the cursor. Its position changes, but its size remains fixed.
B_TRACK_RECT_CORNER The left top corner of the rectangle remains fixed within the view while its right and bottom edges move with the cursor.

Tracking is typically initiated from within a BView's MouseDown() function and is allowed to continue as long as a mouse button is held down. For example:

   void MyView::MouseDown(BPoint point)
   {
       unit32 buttons;
   
       BRect rect(point, point);
       BeginRectTracking(rect, B_TRACK_RECT_CORNER);
       do {
           snooze(30 * 1000);
           GetMouse(&point, &buttons);
       } while ( buttons );
       EndRectTracking();
   
       rect.SetRightBottom(point);
       . . .
   }

This example uses BeginRectTracking() to drag out a rectangle from the point recorded for a mouse-down event. It sets up a modal loop to periodically check on the state of the mouse buttons. Tracking ends when the user releases all buttons. The right and bottom sides of the rectangle are then updated from the cursor location last reported by the GetMouse() function.

See also: ConvertToScreen(), GetMouse()

Bounds()

      BRect Bounds(void) const

Returns the BView's bounds rectangle. If the BView is attached to a window, this function returns the rectangle kept by the Application Server. If not, it returns a rectangle the same size as the BView's frame rectangle, but with the left and top sides at 0.0.

If Bounds() is called during an update, the BView is already aware of the current rectangle (the update message included it); it can therefore return it directly without communicating with the Application Server. However, if called at other times, Bounds() must get the latest rectangle from the server, an interaction that has a certain amount of overhead.

See also: Frame()

ChildAt() see Parent()

ConstrainClippingRegion()

      virtual void ConstrainClippingRegion(BRegion *region)

Restricts the drawing that the BView can do to region.

The Application Server keeps track of a clipping region for each BView that's attached to a window. It clips all drawing the BView does to that region; the BView can't draw outside of it.

By default, the clipping region contains only the visible area of the view and, during an update, only the area that actually needs to be drawn. By passing a region to this function, an application can further restrict the clipping region. When calculating the clipping region, the server intersects it with the region provided. The BView can draw only in areas common to the region passed and the clipping region as the server would otherwise calculate it. The region passed can't expand the clipping region beyond what it otherwise would be.

The clipping region is additionally affected by any items on the state stack. If any saved states contain clipping regions, then the actual clipping region used by the Application Server is the intersection of the local clipping region (as set by this method) and the regions stored on the state stack.

If called during an update, ConstrainClippingRegion() restricts the clipping region only for the duration of the update.

Calls to ConstrainClippingRegion() are not additive; each region that's passed replaces the one that was passed in the previous call. Passing a NULL pointer removes the previous region without replacing it. The function works only for BViews that are attached to a window.

See also: GetClippingRegion(), Draw()

ConvertToParent(), ConvertFromParent()

      BPoint ConvertToParent(BPoint localPoint) const
      void ConvertToParent(BPoint *localPoint) const
      BRect ConvertToParent(BRect localRect) const
      void ConvertToParent(BRect *localRect) const
      BPoint ConvertFromParent(BPoint parentPoint) const
      void ConvertFromParent(BPoint *parentPoint) const
      BRect ConvertFromParent(BRect parentRect) const
      void ConvertFromParent(BRect *parentRect) const

These functions convert points and rectangles to and from the coordinate system of the BView's parent. ConvertToParent() converts localPoint or localRect from the BView's coordinate system to the coordinate system of its parent BView. ConvertFromParent() does the opposite; it converts parentPoint or parentRect from the coordinate system of the BView's parent to the BView's own coordinate system.

If the point or rectangle is passed by value, the function returns the converted value. If a pointer is passed, the conversion is done in place.

Both functions fail if the BView isn't attached to a window.

See also: ConvertToScreen()

ConvertToScreen(), ConvertFromScreen()

      BPoint ConvertToScreen(BPoint localPoint) const
      void ConvertToScreen(BPoint *localPoint) const
      BRect ConvertToScreen(BRect localRect) const
      void ConvertToScreen(BRect *localRect) const
      BPoint ConvertFromScreen(BPoint screenPoint) const
      void ConvertFromScreen(BPoint *screenPoint) const
      BRect ConvertFromScreen(BRect screenRect) const
      void ConvertFromScreen(BRect *screenRect) const

ConvertToScreen() converts localPoint or localRect from the BView's coordinate system to the global screen coordinate system. ConvertFromScreen() makes the opposite conversion; it converts screenPoint or screenRect from the screen coordinate system to the BView's local coordinate system.

If the point or rectangle is passed by value, the function returns the converted value. If a pointer is passed, the conversion is done in place.

The screen coordinate system has its origin, (0.0, 0.0), at the left top corner of the main screen.

Neither function will work if the BView isn't attached to a window.

See also: BWindow::ConvertToScreen(), ConvertToParent()

CopyBits()

      void CopyBits(BRect source, BRect destination)

Copies the image displayed in the source rectangle to the destination rectangle, where both rectangles lie within the view and are stated in the BView's coordinate system.

If the two rectangles aren't the same size, the source image is scaled to fit.

If not all of the destination rectangle lies within the BView's visible region, the source image is clipped rather than scaled.

If not all of the source rectangle lies within the BView's visible region, only the visible portion is copied. It's mapped to the corresponding portion of the destination rectangle. The BView is then invalidated so its Draw() function will be called to update the part of the destination rectangle that can't be filled with the source image.

The BView must be attached to a window.

CountChildren() see Parent()

DetachedFromWindow(), AllDetached()

      virtual void DetachedFromWindow(void)
      virtual void AllDetached(void)

Implemented by derived classes to make any adjustments necessary when the BView is about to be removed from a window's view hierarchy. These two functions parallel the more commonly implemented AttachedToWindow() and AllAttached() functions.

DetachedFromWindow() notifications work their way down the hierarchy of views being detached, followed by AllDetached() notifications, which work their way up the hierarchy. The second function call permits an ancestor view to take actions that depend on calculations a descendant might have to make when it's first notified of being detached.

The BView is still attached to the window when both functions are called.

See also: AttachedToWindow()

DragMessage()

      void DragMessage(BMessage *message, BBitmap *image, BPoint point, 
         BHandler *replyTarget = NULL)
      void DragMessage(BMessage *message, BRect rect, 
         BHandler *replyTarget = NULL)

Initiates a drag-and-drop session.

message, is a BMessage object that bundles the information that will be dragged and dropped on the destination view. The caller retains responsibility for this object and can delete it after DragMessage() return (the BView makes a copy of the message).

image, is a bitmap that the user can drag. The bitmap is automatically freed when the message is dropped.

1 bit-per-pixel bitmaps aren't supported; you should avoid using them.

point locates the hotspot within image (in the bitmap's coordinate system). This is the point that's aligned with the location passed to MouseDown() or returned by GetMouse().

rect defines the dimensions of an outline rectangle that you can instead of a bitmap. The rectangle is stated in the BView's coordinate system.

replyTarget, names the object that you want to handle a message that might be sent in reply to the dragged message. If replyTarget is NULL, as it is by default, any reply that's received will be directed to the BView object that initiated the drag-and-drop session.

This function works only for BViews that are attached to a window.

Draw()

      virtual void Draw(BRect updateRect)

Implemented by derived classes to draw the updateRect portion of the view. The update rectangle is stated in the BView's coordinate system.

Draw() is called as the result of update messages whenever the view needs to present itself on-screen. This may happen when:

The window the view is in is first shown on-screen, or shown after being hidden (see the BWindow version of the Hide() function).
The view is made visible after being hidden (see BView's Hide() function).
Obscured parts of the view are revealed, as when a window is moved from in front of the view or an image is dragged across it.
The view is resized.
The contents of the view are scrolled (see ScrollBy()).
A child view is added, removed, or resized.
A rectangle has been invalidated that includes at least some of the view (see Invalidate()).
CopyBits() can't completely fill a destination rectangle within the view.

Draw() is also called from a BPrintJob object's DrawView() function to draw the view on a printed page. IsPrinting() returns true when the BView is drawing for the printer and false when it's drawing to the screen. When printing, you may want to recalculate layouts, substitute fonts, turn antialiasing off, scale the size of a coordinate unit, or make other adjustments to ensure the quality of the printed image.

When drawing to the screen, the updateRect is the smallest rectangle that encloses the current clipping region for the view. Since the Application Server won't render anything on-screen that's outside the clipping region, an application will be more efficient if it avoids producing drawing instructions for images that don't intersect with the rectangle. For still more efficiency and precision, you can ask for the clipping region itself (by calling GetClippingRegion()) and confine drawing to images that intersect with it.

When printing, the updateRect matches the rectangle passed to DrawView() and may lie outside the clipping region. The clipping region is not enforced for printing, but the Print Server clips the BView's drawing to the specified rectangle.

See also: BWindow::UpdateIfNeeded(), Invalidate(), GetClippingRegion(), BPrintJob::DrawView(), IsPrinting()

DrawBitmap(), DrawBitmapAsync()

      void DrawBitmap(const BBitmap *image)
      void DrawBitmap(const BBitmap *image, BPoint point)
      void DrawBitmap(const BBitmap *image, BRect destination)
      void DrawBitmap(const BBitmap *image, BRect source, BRect destination)
      void DrawBitmapAsync(const BBitmap *image)
      void DrawBitmapAsync(const BBitmap *image, BPoint point)
      void DrawBitmapAsync(const BBitmap *image, BRect destination)
      void DrawBitmapAsync(const BBitmap *image, BRect source, BRect destination)

These functions place a bitmap image in the view at the current pen position, at the point specified, or within the designated destination rectangle. The point and the destination rectangle are stated in the BView's coordinate system.

If a source rectangle is given, only that part of the bitmap image is drawn. Otherwise, the entire bitmap is placed in the view. The source rectangle is stated in the internal coordinates of the BBitmap object.

If the source image is bigger than the destination rectangle, it's scaled to fit.

The two functions differ in only one respect: DrawBitmap() waits for the Application Server to finish rendering the image before it returns. DrawBitmapAsync() doesn't wait; it passes the image to the server and returns immediately. The latter function can be more efficient in some cases--for example, you might use an asynchronous function to draw several bitmaps and then call Sync() to wait for them all to finish rather than wait for each one individually:

   DrawBitmapAsync(bitmapOne, firstPoint);
   DrawBitmapAsync(bitmapTwo, secondPoint);
   DrawBitmapAsync(bitmapThree, thirdPoint);
   Sync();

Or, if you can cram some useful work between the time you send the bitmap to the Application Server and the time you need to be sure that it has appeared on-screen, DrawBitmapAsync() will free your thread to do that work immediately:

   DrawBitmapAsync(someBitmap, somePoint);
   /* do something else */ 
   Sync();

See also: the "Drawing" section near the beginning of this chapter, the BBitmap class, Sync()

DrawChar()

      void DrawChar(char c)
      void DrawChar(char c, BPoint point)

Draws the character c at the current pen position--or at the point specified--and moves the pen to a position immediately to the right of the character. This function is equivalent to passing a string of one character to DrawString(). The point is specified in the BView's coordinate system.

See also: DrawString()

DrawingMode() see SetDrawingMode()

DrawPicture()

      void DrawPicture(const BPicture *picture)
      void DrawPicture(const BPicture *picture, BPoint point) 
      void DrawPicture(const char *filename, off_t offset, BPoint point)

Draws the previously recorded picture at the current pen position--or at the specified point in the BView's coordinate system. The point or pen position is taken as the coordinate origin for all the drawing instructions recorded in the BPicture. The last form of the method plays a picture from an arbitrary offset of a file.

Nothing that's done in the BPicture can affect anything in the BView's graphics state--for example, the BPicture can't reset the current high color or the pen position. Conversely, nothing in the BView's current graphics state affects the drawing instructions captured in the picture. The graphics parameters that were in effect when the picture was recorded determine what the picture looks like.

DrawPicture() is an asynchronous function, in that it initiates drawing and may return before drawing is completed. If it's important that the picture be drawn, and you're going to delete the picture shortly after calling DrawPicture() (either by explicitly deleting it or by exiting the function whose stack the picture resides on), call Sync() before deleting the picture. Otherwise, it might not draw.

See also: BeginPicture(), SetDiskMode(), the BPicture class

DrawString()

      void DrawString(const char *string, 
         escapement_delta *delta = NULL)
      void DrawString(const char *string, int32 length, 
         escapement_delta *delta = NULL)
      void DrawString(const char *string, BPoint point, 
         escapement_delta *delta = NULL)
      void DrawString(const char *string, int32 length, BPoint point, 
         escapement_delta *delta = NULL)

Draws the characters encoded in length bytes of string--or, if the number of bytes isn't specified, all the characters in the string, up to the null terminator ('\\0'). Characters are drawn in the BView's current font. The font's direction determines whether the string is drawn left-to-right or right-to-left. Its rotation determines the angle of the baseline (horizontal for an unrotated font). The spacing mode of the font determines how characters are positioned within the string and the string width.

This function places the characters on a baseline that begins at the current pen position--or at the specified point in the BView's coordinate system. It draws the characters to the right (assuming an unrotated font) and moves the pen to the baseline immediately past the characters drawn. For a left-to-right font, the pen will be in position to draw the next character, as shown below:

The characters are drawn in the opposite direction for a right-to-left font, but the pen still moves left-to-right:

For a font that's read from left-to-right, a series of simple DrawString() calls (with no point specified) will produce a continuous string. For example, these two lines of code,

   DrawString("tog");
   DrawString("ether");

will produce the same result as this one,

   DrawString("together");

except if the spacing mode is B_STRING_SPACING. Under B_STRING_SPACING, character placements are adjusted keeping the string width constant. The adjustments are contextually dependent on the string and may therefore differ depending on whether there are two strings ("tog" and "ether") or just one ("together").

If a delta argument is provided, DrawString() adds the additional amounts specified in the escapement_delta structure to the width of each character. This structure has two fields:

float nonspaceThe amount to add to the width of characters that have visible glyphs (that put ink on the printed page).

float spaceThe amount to add to the width of characters that have escapements, but don't have visible glyphs (characters that affect the position of surrounding characters but don't put ink on the page).

When drawing to the screen, DrawString() uses antialiasing--unless the BView's font disables it or the font size is large enough (over 1,000.0 points) so that its benefits aren't required. Antialiasing produces colors at the margins of character outlines that are intermediate between the color of the text (the BView's high color) and the color of the background against which the text is drawn. When drawing in B_OP_COPY mode, antialiasing requires the BView's low color to match the background color.

It's much faster to draw a string in B_OP_COPY mode than in any other mode. If you draw the same string repeatedly in the same location in B_OP_OVER mode without erasing, antialiasing will produce different, and worse, results each time as the intermediate color it previously produced is treated as the new background each time. Antialiasing doesn't produce pleasing results in B_OP_SELECT mode.

This is a graphical drawing function, so any character that doesn't have an escapement or a visible representation (including white space) is replaced by an undefined character that can be drawn (currently an empty box). This includes all control characters (those with values less than B_SPACE, 0x20).

DrawString() doesn't erase before drawing.

See also: MovePenBy(), SetFontName(), the BFont class

EndLineArray() see BeginLineArray()

EndPicture() see AppendToPicture()

EndRectTracking() see BeginRectTracking()

FillArc() see StrokeArc()

FillBezier() see StrokeBezier()

FillEllipse() see StrokeEllipse()

FillPolygon() see StrokePolygon()

FillRect() see StrokeRect()

FillRegion()

      void FillRegion(BRegion *region, pattern aPattern = B_SOLID_HIGH) const

Fills the region with the pattern specified by aPattern--or, if no pattern is specified, with the current high color. Filling a region is equivalent to filling all the rectangles that define the region.

See also: the BRegion class

FillRoundRect() see StrokeRoundRect()

FillTriangle() see StrokeTriangle()

FindView()

      BView *FindView(const char *name) const

Returns the BView identified by name, or NULL if the view can't be found. Names are assigned by the BView constructor and can be modified by the SetName() function inherited from BHandler.

FindView() begins the search by checking whether the BView's name matches name. If not, it continues to search down the view hierarchy, among the BView's children and more distant descendants. To search the entire view hierarchy, use the BWindow version of this function.

See also: BWindow::FindView(), BHandler::SetName()

Flags() see SetFlags()

Flush(), Sync()

      void Flush(void) const
      void Sync(void) const

These functions flush the window's connection to the Application Server. If the BView isn't attached to a window, Flush() does nothing.

If the BView isn't attached to a window, Sync() will crash the application.

For reasons of efficiency, the window's connection to the Application Server is buffered. Drawing instructions destined for the server are placed in the buffer and dispatched as a group when the buffer becomes full. Flushing empties the buffer, sending whatever it contains to the server, even if it's not yet full.

The buffer is automatically flushed on every update. However, if you do any drawing outside the update mechanism--in response to interface messages, for example--you need to explicitly flush the connection so that drawing instructions won't languish in the buffer while waiting for it to fill up or for the next update. You should also flush it if you call any drawing functions from outside the window's thread.

Flush() simply flushes the buffer and returns. It does the same work as BWindow's function of the same name.

Sync() flushes the connection, then waits until the server has executed the last instruction that was in the buffer before returning. This alternative to Flush() prevents the application from getting ahead of the server (ahead of what the user sees on-screen) and keeps both processes synchronized.

It's a good idea, for example, to call Sync(), rather than Flush(), after employing BViews to produce a bitmap image (a BBitmap object). Sync() is the only way you can be sure the image has been completely rendered before you attempt to draw with it.

(Note that all BViews attached to a window share the same connection to the Application Server. Calling Flush() or Sync() for any one of them flushes the buffer for all of them.)

See also: BWindow::Flush(), the BBitmap class

Frame()

      BRect Frame(void) const

Returns the BView's frame rectangle. The frame rectangle is first set by the BView constructor and is altered only when the view is moved or resized. It's stated in the coordinate system of the BView's parent.

If the BView is not attached to a window, Frame() reports the object's own cached conception of its frame rectangle. If it is attached, Frame() reports the Application Server's conception of the rectangle. When a BView is added to a window, its cached rectangle is communicated to the server. While it remains attached, the functions that move and resize the frame rectangle affect the server's conception of the view, but don't alter the rectangle kept by the object. Therefore, if the BView is removed from the window, Frame() will again report the frame rectangle that it had before it was attached, no matter how much it was moved and resized while it belonged to the window.

See also: MoveBy(), ResizeBy(), the BView constructor

FrameMoved()

      virtual void FrameMoved(BPoint parentPoint)

Implemented by derived classes to respond to a notification that the view has moved within its parent's coordinate system. parentPoint gives the new location of the left top corner of the BView's frame rectangle.

FrameMoved() is called only if the B_FRAME_EVENTS flag is set and the BView is attached to a window.

If the view is both moved and resized, FrameMoved() is called before FrameResized(). This might happen, for example, if the BView's automatic resizing mode is a combination of B_FOLLOW_TOP_BOTTOM and B_FOLLOW_RIGHT and its parent is resized both horizontally and vertically.

The default (BView) version of this function is empty.

Currently, FrameMoved() is also called when a hidden window is shown on-screen.

See also: MoveBy(), BWindow::FrameMoved(), SetFlags()

FrameResized()

      virtual void FrameResized(float width, float height)

Implemented by derived classes to respond to a notification that the view has been resized. The arguments state the new width and height of the view. The resizing could have been the result of a user action (resizing the window) or of a programmatic one (calling ResizeTo() or ResizeBy()).

FrameResized() is called only if the B_FRAME_EVENTS flag is set and the BView is attached to a window.

BView's version of this function is empty.

See also: ResizeBy(), BWindow::FrameResized(), SetFlags()

GetClippingRegion()

      void GetClippingRegion(BRegion *region) const

Modifies the BRegion object passed as an argument so that it describes the current local clipping region of the BView, the region where the BView is allowed to draw. It's most efficient to allocate temporary BRegions on the stack:

   BRegion clipper;
   GetClippingRegion(&clipper);
   . . .

Ordinarily, the clipping region is the same as the visible region of the view, the part of the view currently visible on-screen. The visible region is equal to the view's bounds rectangle minus:

The frame rectangles of its children,
Any areas that are clipped because the view doesn't lie wholly within the frame rectangles of all its ancestors in the view hierarchy, and
Any areas that are obscured by other windows or that lie in a part of the window that's off-screen.

The clipping region can be smaller than the visible region if the program restricted it by calling ConstrainClippingRegion(). It will exclude any area that doesn't intersect with the region passed to ConstrainClippingRegion().

The clipping region is additionally modified by any items on the state stack. If any saved states contain clipping regions, then the actual clipping region used by the Application Server is the intersection of the local clipping region (as set by ConstrainClippingRegion()) and the regions stored on the state stack.

While the BView is being updated, the clipping region contains just those parts of the view that need to be redrawn. This may be smaller than the visible region, or the region restricted by ConstrainClippingRegion(), if:

The update occurs during scrolling. The clipping region will exclude any of the view's visible contents that the Application Server is able to shift to their new location and redraw automatically.
The view rectangle has grown (because, for example, the user resized the window larger) and the update is needed only to draw the new parts of the view.
The update was caused by Invalidate() and the rectangle passed to Invalidate() didn't cover all of the visible region.
The update was necessary because CopyBits() couldn't fill all of a destination rectangle.

This function works only if the BView is attached to a window. Unattached BViews can't draw and therefore have no clipping region.

See also: ConstrainClippingRegion(), Draw(), Invalidate()

GetFont() see SetFont()

GetFontHeight()

      void GetFontHeight(font_height *fontHeight) const

Gets the height of the BView's font. This function provides the same information as BFont's GetHeight(). The following code

   font_height height;
   myView->GetFontHeight(&height);

is equivalent to:

   font_height height;
   BFont font;
   myView->GetFont(&font);
   font.GetHeight(&height);

See the BFont class for more information.

See also: BFont::GetHeight()

GetMouse()

      void GetMouse(BPoint *cursor, uint32 *buttons, bool checkQueue = true)

Provides the location of the cursor and the state of the mouse buttons. The position of the cursor is recorded in the variable referred to by cursor; it's provided in the BView's own coordinates. A bit is set in the variable referred to by buttons for each mouse button that's down. This mask may be 0 (if no buttons are down) or it may contain one or more of the following constants:

B_PRIMARY_MOUSE_BUTTON
B_SECONDARY_MOUSE_BUTTON
B_TERTIARY_MOUSE_BUTTON

The cursor doesn't have to be located within the view for this function to work; it can be anywhere on-screen. However, the BView must be attached to a window.

If the checkQueue flag is set to false, GetMouse() provides information about the current state of the mouse buttons and the current location of the cursor.

If checkQueue is true, as it is by default, this function first looks in the message queue for any pending reports of mouse-moved or mouse-up events. If it finds any, it takes the one that has been in the queue the longest (the oldest message), removes it from the queue, and reports the cursor location and button states that were recorded in the message. Each GetMouse() call removes another message from the queue. If the queue doesn't hold any B_MOUSE_MOVED or B_MOUSE_UP messages, GetMouse() reports the current state of the mouse and cursor, just as if checkQueue were false.

This function is typically called from within a MouseDown() function to track the location of the cursor and wait for the mouse button to go up. By having it check the message queue, you can be sure that you haven't overlooked any of the cursor's movement or missed a mouse-up event (quickly followed by another mouse-down) that might have occurred before the first GetMouse() call.

See also: modifiers()

GetPreferredSize(), ResizeToPreferred()

      virtual  void GetPreferredSize(float *width, float *height) 
      virtual void ResizeToPreferred(void)

GetPreferredSize() is implemented by derived classes to write the preferred width and height of the view into the variables the width and height arguments refer to. Derived classes generally make this calculation based on the view's contents. For example, a BButton object reports the optimal size for displaying the button border and label given the current font.

ResizeToPreferred() resizes the BView's frame rectangle to the preferred size, keeping its left and top sides constant.

See also: ResizeTo()

GetStringWidths() see StringWidth()

GetSupportedSuites()

      virtual status_t GetSupportedSuites(BMessage *message)

Adds the name "suite/vnd.Be-view" to the message. See 23854: head2: Scripting Support in the class overview for more information.

See also: BHandler::GetSupportedSuites()

Hide(), Show()

      virtual void Hide(void)
      virtual void Show(void)

These functions hide a view and show it again.

Hide() makes the view invisible without removing it from the view hierarchy. The visible region of the view will be empty and the BView won't receive update messages. If the BView has children, they also are hidden.

Show() unhides a view that had been hidden. This function doesn't guarantee that the view will be visible to the user; it merely undoes the effects of Hide(). If the view didn't have any visible area before being hidden, it won't have any after being shown again (given the same conditions).

Calls to Hide() and Show() can be nested. For a hidden view to become visible again, the number of Hide() calls must be matched by an equal number of Show() calls.

However, Show() can only undo a previous Hide() call on the same view. If the view became hidden when Hide() was called to hide the window it's in or to hide one of its ancestors in the view hierarchy, calling Show() on the view will have no effect. For a view to come out of hiding, its window and all its ancestor views must be unhidden.

Hide() and Show() can affect a view before it's attached to a window. The view will reflect its proper state (hidden or not) when it becomes attached. Views are created in an unhidden state.

See also: BWindow::Hide(), IsHidden()

HighColor() see SetHighColor()

Invalidate()

      void Invalidate(BRect rect)
      void Invalidate(void)

Invalidates the rect portion of the view, causing update messages--and consequently Draw() notifications--to be generated for the BView and all descendants that lie wholly or partially within the rectangle. The rectangle is stated in the BView's coordinate system.

If no rectangle is specified, the BView's entire bounds rectangle is invalidated.

Since only BViews that are attached to a window can draw, only attached BViews can be invalidated.

See also: Draw(), GetClippingRegion(), BWindow::UpdateIfNeeded()

InvertRect()

      void InvertRect(BRect rect)

Inverts all the colors displayed within the rect rectangle. A subsequent InvertRect() call on the same rectangle restores the original colors.

The rectangle is stated in the BView's coordinate system.

See also: BScreen::ColorMap

IsFocus()

      bool IsFocus(void) const

Returns true if the BView is the current focus view for its window, and false if it's not. The focus view changes as the user chooses one view to work in and then another--for example, as the user moves from one text field to another when filling out an on-screen form. The change is made programmatically through the MakeFocus() function.

See also: BWindow::CurrentFocus(), MakeFocus()

IsHidden()

      bool IsHidden(void) const

Returns true if the view has been hidden by the Hide() function, and false otherwise.

This function returns true whether Hide() was called to hide the BView itself, to hide an ancestor view, or to hide the BView's window. When a window is hidden, all its views are hidden with it. When a BView is hidden, all its descendants are hidden with it.

If the view has no visible region--perhaps because it lies outside its parent's frame rectangle or is obscured by a window in front--this function may nevertheless return false. It reports only whether the Hide() function has been called to hide the view, hide one of the view's ancestors in the view hierarchy, or hide the window where the view is located.

If the BView isn't attached to a window, IsHidden() returns the state that it will assume when it becomes attached. By default, views are not hidden.

See also: Hide()

IsPrinting()

      bool IsPrinting(void) const

Returns true if the BView is being asked to draw for the printer, and false if the drawing it produces will be rendered on-screen (or if the BView isn't being asked to draw at all).

This function is typically called from within Draw() to determine whether the drawing it does is destined for the printer or the screen. When drawing to the printer, the BView may choose different parameters--such as fonts, bitmap images, or colors--than when drawing to the screen.

See also: the BPrintJob class, Draw()

KeyDown()

      virtual void KeyDown(const char *bytes, int32 numBytes)

Implemented by derived classes to respond to a B_KEY_DOWN message reporting keyboard input. Whenever a BView is the focus view of the active window, it receives a KeyDown() notification for each character the user types, except for those that:

Are produced while a Command key is held down. Command key events are interpreted as keyboard shortcuts.
Are produced by the Tab key when an Option key is held down. Option-Tab events are invariably interpreted as instructions to change the focus view (for keyboard navigation); they work even where Tab alone does not.
Can operate the default button in a window. The BButton object's KeyDown() function is called, rather than the focus view's.

The first argument, bytes, is an array that encodes the character mapped to the key the user pressed. The second argument, numBytes, tells how many bytes are in the array; there will always be at least one. The bytes value follows the character encoding of the BView's font. Typically, the encoding is Unicode UTF-8 (B_UNICODE_UTF8), so there may be more than one byte per character. The bytes array is not null-terminated; '\\0' is a valid character value.

The character value takes into account any modifier keys that were held down or keyboard locks that were on at the time of the keystroke. For example, Shift-i is reported as uppercase 'I' (0x49) and Control-i is reported as a B_TAB (0x09).

Single-byte characters can be tested against ASCII codes and these constants:

B_BACKSPACE B_LEFT_ARROW B_INSERT
B_ENTER B_RIGHT_ARROW B_DELETE
B_RETURN B_UP_ARROW B_HOME
B_SPACE B_DOWN_ARROW B_END
B_TAB B_PAGE_UP
B_ESCAPE B_FUNCTION_KEY B_PAGE_DOWN

B_ENTER and B_RETURN are the same character, a newline ('\\n').

Only keys that generate characters produce key-down events; the modifier keys on their own do not.

You can determine which modifier keys were being held down at the time of the event by calling BLooper's CurrentMessage() function and looking up the "modifiers" entry in the BMessage it returns. If the bytes character is B_FUNCTION_KEY and you want to know which key produced the character, you can look up the "key" entry in the BMessage and test it against these constants:

B_F1_KEY B_F6_KEY B_F11_KEY
B_F2_KEY B_F7_KEY B_F12_KEY
B_F3_KEY B_F8_KEY B_PRINT_KEY (Print Screen)
B_F4_KEY B_F9_KEY B_SCROLL_KEY (Scroll Lock)
B_F5_KEY B_F10_KEY B_PAUSE_KEY

For example:

   if ( bytes[0] == B_FUNCTION_KEY ) {
       BMessage *msg = Window()->CurrentMessage();
       if ( msg ) {
           int32 key;
           msg->FindInt32("key", &key);
           switch ( key ) {
           case B_F1_KEY:
               . . .
               break;
           case B_F2_KEY:
               . . .
               break;
            . . .
           }
       }
   }

The BView version of KeyDown() handles keyboard navigation from view to view through B_TAB characters. If the view you define is navigable, its KeyDown() function should permit B_SPACE characters to operate the object and perhaps allow the arrow keys to navigate inside the view. It should also call the inherited version of KeyDown() to enable between-view navigation. For example:

   void MyView::KeyDown(const char *bytes, int32 numBytes)
   {
       if ( numBytes == 1 ) {
           switch ( bytes[0] ) {
           case B_SPACE:
               /* mimic a click in the view */
               break;
           case B_RIGHT_ARROW:
               /* move one position to the right in the view */
               break;
           case B_LEFT_ARROW:
               /* move one position to the left in the view */
               break;
           default:
               baseClass::KeyDown(bytes, numBytes);
               break;
           }
       }
   }

If your BView is navigable but needs to respond to B_TAB characters--for example, if it permits users to insert tabs in a text string--its KeyDown() function should simply grab the characters and not pass them to the inherited function. Users will have to rely on the Option-Tab combination to navigate from your view.

See also: the Keyboard Information appendix, "B_KEY_DOWN" in the Message Protocols appendix, BWindow::SetDefaultButton(), modifiers()

KeyUp()

      virtual void KeyUp(const char *bytes, int32 numBytes)

Implemented by derived classes to respond to a B_KEY_UP message reporting that the user released a key on the keyboard. The same set of keys that produce B_KEY_DOWN messages when they're pressed produce B_KEY_UP messages when they're released. The bytes and numBytes arguments encode the character mapped to the key the user released; they work exactly like the same arguments passed to KeyDown().

Some B_KEY_DOWN messages are swallowed by the system and are never dispatched by calling KeyDown(); others are dispatched, but not to the focus view. In contrast, all B_KEY_UP messages are dispatched by calling KeyUp() for the focus view of the active window. Since the focus view and active window can change between the time a key is pressed and the time it's released, this may or may not be the same BView that was notified of the B_KEY_DOWN message.

See also: KeyDown(), "B_KEY_DOWN" in the Message Protocols appendix

LeftTop()

      BPoint LeftTop(void) const

Returns the coordinates of the left top corner of the view--the smallest x and y coordinate values within the bounds rectangle.

See also: BRect::LeftTop(), Bounds()

LineCapMode() see SetLineMode()

LineJoinMode() see SetLineMode()

LineMiterLimit() see SetLineMode()

LowColor() see SetHighColor()

MakeFocus()

      virtual void MakeFocus(bool focused = true)

Makes the BView the current focus view for its window (if the focused flag is true), or causes it to give up that status (if focused is false). The focus view is the view that displays the current selection and is expected to handle reports of key-down events when the window is the active window. There can be no more than one focus view per window at a time.

When called to make a BView the focus view, this function invokes MakeFocus() for the previous focus view, passing it an argument of false. It's thus called twice--once for the new and once for the old focus view.

Calling MakeFocus() is the only way to make a view the focus view; the focus doesn't automatically change on mouse-down events. BViews that can display the current selection (including an insertion point) or that can accept pasted data should call MakeFocus() in their MouseDown() functions.

A derived class can override MakeFocus() to add code that takes note of the change in status. For example, a BView that displays selectable data may want to highlight the current selection when it becomes the focus view, and remove the highlighting when it's no longer the focus view. A BView that participates in the keyboard navigation system should visually indicate that it can be operated from the keyboard when it becomes the focus view, and remove that indication when the user navigates to another view and it's notified that it's no longer the focus view.

If the BView isn't attached to a window, this function has no effect.

See also: BWindow::CurrentFocus(), IsFocus()

MessageReceived()

      virtual void MessageReceived(BMessage *message)

Augments the BHandler version of MessageReceived() to handle scripting messages for the BView.

See also: BHandler::MessageReceived()

MouseDown()

      virtual void MouseDown(BPoint point)

Implemented by derived classes to respond to a message reporting a mouse-down event within the view. The location of the cursor at the time of the event is given by point in the BView's coordinates.

MouseDown() functions are often implemented to track the cursor while the user holds the mouse button down and then respond when the button goes up. You can call the GetMouse() function to learn the current location of the cursor and the state of the mouse buttons. For example:

   void MyView::MouseDown(BPoint point)
   {
       uint32 buttons = 0;
       . . .
       Window()->CurrentMessage()->FindInt32("buttons", &buttons);
       while ( buttons ) {
           . . .
           snooze(20 * 1000);
           GetMouse(&point, &buttons, true);
       }
       . . .
   }

It's important to snooze between GetMouse() calls so that the loop doesn't monopolize system resources; 20,000 microseconds is a minimum time to wait.

To get complete information about the mouse-down event, look inside the BMessage object returned by BLooper's CurrentMessage() function. The "clicks" entry in the message can tell you if this mouse-down is a solitary event or the latest in a series constituting a multiple click.

The BView version of MouseDown() is empty.

See also: "B_MOUSE_DOWN" in the Message Protocols appendix, GetMouse()

MouseMoved()

      virtual void MouseMoved(BPoint point, uint32 transit, const BMessage *message)

Implemented by derived classes to respond to reports of mouse-moved events associated with the view. As the user moves the cursor over a window, the Application Server generates a continuous stream of messages reporting where the cursor is located.

The first argument, point, gives the cursor's new location in the BView's coordinate system. The second argument, transit, is one of three constants,

B_ENTERED_VIEW
B_INSIDE_VIEW
B_EXITED_VIEW

which explains whether the cursor has just entered the visible region of the view, is now inside the visible region having previously entered, or has just exited from the view. When the cursor crosses a boundary separating the visible regions of two views (perhaps moving from a parent to a child view, or from a child to a parent), MouseMoved() is called for each of the BViews, once with a transit code of B_EXITED_VIEW and once with a code of B_ENTERED_VIEW.

If the user is dragging a bundle of information from one location to another, the final argument, message, is a pointer to the BMessage object that holds the information. If a message isn't being dragged, message is NULL.

A MouseMoved() function might be implemented to ignore the B_INSIDE_VIEW case and respond only when the cursor enters or exits the view. For example, a BView might alter its display to indicate whether or not it can accept a message that has been dragged to it. Or it might be implemented to change the cursor image when it's over the view.

MouseMoved() notifications should not be used to track the cursor inside a view. Use the GetMouse() function instead. GetMouse() provides the current cursor location plus information on whether any of the mouse buttons are being held down.

The default version of MouseMoved() is empty.

See also: "B_MOUSE_MOVED" in the Message Protocols appendix, DragMessage()

MouseUp()

Currently unimplemented. Use GetMouse() to detect the state of the mouse.

MoveBy(), MoveTo()

      void MoveBy(float horizontal, float vertical)
      void MoveTo(BPoint point)
      void MoveTo(float x, float y)

These functions move the view in its parent's coordinate system without altering its size.

MoveBy() adds horizontal coordinate units to the left and right components of the frame rectangle and vertical units to the top and bottom components. If horizontal and vertical are positive, the view moves downward and to the right. If they're negative, it moves upward and to the left.

MoveTo() moves the upper left corner of the view to point--or to (x, y)--in the parent view's coordinate system and adjusts all coordinates in the frame rectangle accordingly.

Neither function alters the BView's bounds rectangle or coordinate system.

None of the values passed to these functions should specify fractional coordinates; the sides of a view must line up on screen pixels. Fractional values will be rounded to the closest whole number.

If the BView is attached to a window, these functions cause its parent view to be updated, so the BView is immediately displayed in its new location. If it doesn't have a parent or isn't attached to a window, these functions merely alter its frame rectangle.

See also: FrameMoved(), ResizeBy(), Frame()

MovePenBy(), MovePenTo(), PenLocation()

      void MovePenBy(float horizontal, float vertical)
      void MovePenTo(BPoint point)
      void MovePenTo(float x, float y)
      BPoint PenLocation(void) const

These functions move the pen (without drawing a line) and report the current pen location.

MovePenBy() moves the pen horizontal coordinate units to the right and vertical units downward. If horizontal or vertical are negative, the pen moves in the opposite direction. MovePenTo() moves the pen to point--or to (x, y)--in the BView's coordinate system.

Some drawing functions also move the pen--to the end of whatever they draw. In particular, this is true of StrokeLine(), DrawString(), and DrawChar(). Functions that stroke a closed shape (such as StrokeEllipse()) don't move the pen.

The pen location is a parameter of the BView's graphics environment, which is maintained by both the Application Server and the BView. If the BView doesn't belong to a window, MovePenTo() and MovePenBy() cache the location, so that later, when the BView becomes attached to a window, it can be handed to the server to become the operable pen location for the BView. If the BView belongs to a window, these functions alter both the server parameter and the client-side cache.

PenLocation() returns the point where the pen is currently positioned in the BView's coordinate system. Because of the cache, this shouldn't entail contacting the server. The default pen position is (0.0, 0.0).

See also: SetPenSize()

MoveTo() see MoveBy()

NextSibling() see Parent()

Origin() see SetOrigin()

Parent(), NextSibling(), PreviousSibling(), ChildAt(), CountChildren()

      BView *Parent(void) const
      BView *NextSibling(void) const
      BView *PreviousSibling(void) const
      BView *ChildAt(int32 index) const
      int32 CountChildren(void) const

These functions provide various ways of navigating the view hierarchy. Parent() returns the BView's parent view, unless the parent is the top view of the window, in which case it returns NULL. It also returns NULL if the BView doesn't belong to a view hierarchy and has no parent.

All the children of the same parent are arranged in a linked list. NextSibling() returns the next sibling of the BView in the list, or NULL if the BView is the last child of its parent. PreviousSibling() returns the previous sibling of the BView, or NULL if the BView is the first child of its parent.

ChildAt() returns the view at index in the list of the BView's children, or NULL if the BView has no such child. Indices begin at 0 and there are no gaps in the list. CountChildren() returns the number of children the BView has. If the BView has no children, CountChildren() returns NULL, as will ChildAt() for all indices, including 0.

To scan the list of a BView's children, you can increment the index passed to ChildAt() until it returns NULL. However, it's more efficient to ask for the first child and then use NextSibling() to walk down the sibling list. For example:

   BView *child;
   if ( child = myView->ChildAt(0) ) {
       while ( child ) {
           . . .
           child = child->NextSibling();
       }
   }

See also: AddChild()

PenLocation() see MovePenBy()

PenSize() see SetPenSize()

PopState() see PushState()

PreviousSibling() see Parent()

Pulse()

      virtual void Pulse(void)

Implemented by derived classes to do something at regular intervals. Pulses are regularly timed events, like the tick of a clock or the beat of a steady pulse. A BView receives Pulse() notifications when no other messages are pending, but only if it asks for them with the B_PULSE_NEEDED flag.

The interval between Pulse() calls can be set with BWindow's SetPulseRate() function. The default interval is around 500 milliseconds. The pulse rate is the same for all views within a window, but can vary between windows.

Derived classes can implement a Pulse() function to do something that must be repeated continuously. However, for time-critical actions, you should implement your own timing mechanism.

The BView version of this function is empty.

See also: SetFlags(), the BView constructor, BWindow::SetPulseRate()

PushState(), PopState()

      void PushState(void)
      void PopState(void)

Saves and restores the state from the state stack. A state consists of the following: local and global origins, local and global scales, drawing mode, line cap and join modes, miter limit, pen size and location, foreground and background color, stipple pattern, local and global clipping regions, and the font context. When a state is saved to the stack, a new state context is created, with a local scale of zero, a local origin at (0,0), and no clipping region.

If the BView isn't attached to a window, these functions will crash the application.

RemoveChild() see AddChild()

RemoveSelf()

      bool RemoveSelf(void)

Removes the BView from its parent and returns true, or returns false if the BView doesn't have a parent or for some reason can't be removed from the view hierarchy.

This function acts just like RemoveChild(), except that it removes the BView itself rather than one of its children.

See also: AddChild()

ResizeBy(), ResizeTo()

      void ResizeBy(float horizontal, float vertical)
      void ResizeTo(float width, float height)

These functions resize the view, without moving its left and top sides. ResizeBy() adds horizontal coordinate units to the width of the view and vertical units to the height. ResizeTo() makes the view width units wide and height units high. Both functions adjust the right and bottom components of the frame rectangle accordingly.

Since a BView's frame rectangle must be aligned on screen pixels, only integral values should be passed to these functions. Values with fractional components will be rounded to the nearest whole integer.

If the BView is attached to a window, these functions cause its parent view to be updated, so the BView is immediately displayed in its new size. If it doesn't have a parent or isn't attached to a window, these functions merely alter its frame and bounds rectangles.

If the view isn't attached to a window, its frame and bounds rectangles are adjusted, but its children, if any, don't get corresponding adjustments.

See also: FrameResized(), MoveBy(), BRect::Width(), Frame()

ResizeToPreferred() see GetPreferredSize()

ResizingMode() see SetResizingMode()

ResolveSpecifier()

      virtual BHandler *ResolveSpecifier(BMessage *message, int32 index, BMessage *specifier, int32 command, const char *property)

Resolves specifiers for the "Frame", "Hidden", and "View" properties. See 23854: head2: Scripting Support in the class overview and "Scripting" in The Application Kit chapter for more information.

See also: BHandler::ResolveSpecifier()

ScrollBar()

      BScrollBar *ScrollBar(orientation posture) const

Returns a BScrollBar object that scrolls the BView (that has the BView as its target). The requested scroll bar has the posture orientation--B_VERTICAL or B_HORIZONTAL. If the BView isn't the target of a scroll bar with the specified orientation, this function returns NULL.

See also: ScrollBar::SetTarget()

ScrollBy(), ScrollTo()

      void ScrollBy(float horizontal, float vertical)
      virtual void ScrollTo(BPoint point)
      inline void ScrollTo(float x, float y)

These functions scroll the contents of the view, provided that the BView is attached to a window.

ScrollBy() adds horizontal to the left and right components of the BView's bounds rectangle, and vertical to the top and bottom components. This serves to shift the display horizontal coordinate units to the left and vertical units upward. If horizontal and vertical are negative, the display shifts in the opposite direction.

ScrollTo() shifts the contents of the view as much as necessary to put point--or (x, y)--at the upper left corner of its bounds rectangle. The point is specified in the BView's coordinate system.

Anything in the view that was visible before scrolling and also visible afterwards is automatically redisplayed at its new location. The remainder of the view is invalidated, so the BView's Draw() function will be called to fill in those parts of the display that were previously invisible. The update rectangle passed to Draw() will be the smallest possible rectangle that encloses just these new areas. If the view is scrolled in only one direction, the update rectangle will be exactly the area that needs to be drawn.

If the BView is the target of scroll bars, ScrollBy() and ScrollTo() notify the BScrollBar objects of the change in the display so they can update themselves to match. If the contents were scrolled horizontally, they call the horizontal BScrollBar's SetValue() function and pass it the new value of the left side of the bounds rectangle. If they were scrolled vertically, they call SetValue() for the vertical BScrollBar and pass it the new value of the top of the bounds rectangle.

The inline version of ScrollTo() works by creating a BPoint object and passing it to the version that's declared virtual. Therefore, if you want to override either function, you should override the virtual version. (However, due to the peculiarities of C++, overriding any version of an overloaded function hides all versions of the function. For continued access to the nonvirtual version without explicitly specifying the "BView::" prefix, simply copy the inline code from interface/View.h into the derived class.)

See also: GetClippingRegion(), BScrollBar::SetValue()

SetDiskMode()

      void SetDiskMode(const char *filename, off_t offset)

Begins recording a picture to the file with the given filename at the given offset. Subsequent drawing commands sent to the view will be written to the file until EndPicture() is called. The stored commands may be played from the file with DrawPicture().

See also: DrawPicture(), EndPicture()

SetDrawingMode(), DrawingMode()

      virtual void SetDrawingMode(drawing_mode mode)
      drawing_mode DrawingMode(void) const

These functions set and return the BView's drawing mode, which can be any of the following ten constants:

B_OP_COPY B_OP_MIN
B_OP_OVER B_OP_MAX
B_OP_ERASE B_OP_ADD
B_OP_INVERT B_OP_SUBTRACT
B_OP_SELECT B_OP_BLEND

The drawing mode is an element of the BView's graphics environment, which both the Application Server and the BView keep track of. If the BView isn't attached to a window, SetDrawingMode() caches the mode. When the BView is placed in a window and becomes known to the server, the cached value is automatically set as the current drawing mode. If the BView belongs to a window, SetDrawingMode() makes the change in both the server and the cache.

DrawingMode() returns the current mode. Because of the cache, this generally doesn't entail a trip to the server.

The default drawing mode is B_OP_COPY. It and the other modes are explained under "Drawing Modes" in the "Drawing" section of this chapter.

SetFlags(), Flags()

      virtual void SetFlags(uint32 mask)
      uint32 Flags(void) const

These functions set and return the flags that inform the Application Server about the kinds of notifications the BView should receive. The mask set by SetFlags() and the return value of Flags() is formed from combinations of the following constants:

B_WILL_DRAW
B_FULL_UPDATE_ON_RESIZE
B_FRAME_EVENTS
B_PULSE_NEEDED
B_NAVIGABLE
B_NAVIGABLE_JUMP
B_SUBPIXEL_PRECISE

The flags are first set when the BView is constructed; they're explained in the description of the BView constructor. The mask can be 0.

To set just one of the flags, combine it with the current setting:

   myView->SetFlags(Flags() | B_FRAME_EVENTS);

See also: the BView constructor, SetResizingMode()

SetFont(), GetFont()

      virtual void SetFont(const BFont *font, uint32 properties = B_FONT_ALL)
      void GetFont(BFont *font)

SetFont() sets the BView's current font so that it matches the specified properties of the font BFont object. The properties mask is formed by combining the following constants:

B_FONT_FAMILY_AND_STYLE B_FONT_SPACING
B_FONT_SIZE B_FONT_ENCODING
B_FONT_SHEAR B_FONT_FACE
B_FONT_ROTATION B_FONT_FLAGS

Each constant corresponds to a settable property of the BFont object. The default mask, B_FONT_ALL, is a shorthand for all the properties (including any that might be added in future releases). If the mask is 0, SetFont() won't set the BView's font.

GetFont() copies the BView's current font to the BFont object passed as an argument. Modifying this copy doesn't modify the BView's font; it takes an explicit SetFont() call to affect the BView.

For example, this code changes the size of a BView's font and turns antialiasing off:

   BFont font;
   myView->GetFont(&font);
   font.SetSize(67.0);
   font.SetFlags(B_DISABLE_ANTIALIASING);
   myView->SetFont(&font, B_FONT_SIZE | B_FONT_FLAGS);

Since the BFont object that this example code alters is a copy of the BView's current font, it's not strictly necessary to name the properties that are different when calling SetFont(). However, it's more efficient and better practice to do so.

The font is part of the BView's graphic environment. Like other elements in the environment, it can be set whether or not the BView is attached to the window. Graphics parameters are kept by the Application Server and also cached by the BView object.

See also: the BFont class, get_font_family()

SetFontSize()

      void SetFontSize(float points)

Sets the size of the BView's font to points. This function is a shorthand for a SetFont() call that just alters the font size. For example, this line of code

   myView->SetFontSize(12.5);

does the same thing as:

   BFont font;
   font.SetSize(12.5);
   myView->SetFont(&font, B_FONT_SIZE);

See also: the BFont class, SetFont()

SetHighColor(), HighColor(), SetLowColor(), LowColor()

      virtual void SetHighColor(rgb_color color)
      inline void SetHighColor(uchar red, uchar green, uchar blue, uchar alpha = 0)
      rgb_color HighColor(void) const
      virtual void SetLowColor(rgb_color color)
      inline void SetLowColor(uchar red, uchar green, uchar blue, uchar alpha = 0)
      rgb_color LowColor(void) const

These functions set and return the current high and low colors of the BView. These colors combine to form a pattern that's passed as an argument to the Stroke...() and Fill...() drawing functions. The B_SOLID_HIGH pattern is the high color alone, and B_SOLID_LOW is the low color alone.

The default high color is black--red, green, and blue values all equal to 0. The default low color is white--red, green, and blue values all equal to 255.

The alpha component of the color is currently ignored.

The inline versions of SetHighColor() and SetLowColor() take separate arguments for the red, blue, and green color components; they work by creating an rgb_color data structure and passing it to the corresponding function that's declared virtual. Therefore, if you want to override either function, you should override the virtual version. (However, due to the peculiarities of C++, overriding any version of an overloaded function hides all versions of the function. For continued access to the nonvirtual version without explicitly specifying the "BView::" prefix, simply copy the inline code from interface/View.h into the derived class.)

The high and low colors are parameters of the BView's graphics environment, which is kept in the BView's shadow counterpart in the Application Server and cached in the BView. If the BView isn't attached to a window, SetHighColor() and SetLowColor() cache the color value so that later, when the BView is placed in a window and becomes known to the server, the cached value can automatically be registered as the current high or low color for the view. If the BView belongs to a window, these functions alter both the client-side and the server-side values.

HighColor() and LowColor() return the BView's current high and low colors. Because of the cache, this shouldn't entail contacting the Application Server.

See also: "Patterns" in the "Drawing" section of this chapter, SetViewColor()

SetLineMode(), LineJoinMode(), LineCapMode(), LineMiterLimit()

      void SetLineMode(cap_mode lineCap, join_mode lineJoin, 
         float miterLimit = B_DEFAULT_MITER_LIMIT)
      cap_mode LineCapMode(void) const
      join_mode LineJoinMode(void) const
      float LineMiterLimit(void) const

These methods implement support for PostScript-style line cap and join modes. The cap mode determines the shape of the endpoints of stroked paths, while the join mode determines the shape of the corners of the paths (i.e. where two lines meet).

The following values of cap_mode are defined:

B_ROUND_CAP A semicircle is drawn around the endpoint. Its diameter is equal to the width of the line.
B_BUTT_CAP The line is squared off and does not extend beyond the endpoint.
B_SQUARE_CAP The line is squared off, extending past the endpoint for a distance equal to half the width of the line.

Additionally, the following values of join_mode are defined:

B_ROUND_JOIN Acts identically to B_ROUND_CAP, except applied to joins.
B_MITER_JOIN The lines are extended until they touch. If they meet at an angle greater than 2*arcsin(1/miterLimit), a bevel join is used instead.
B_BEVEL_JOIN Butt end caps are used at the common endpoint and the empty area between the caps is filled with a triangle.
B_BUTT_JOIN Acts identically to B_BUTT_CAP, except applied to joins.
B_SQUARE_JOIN Acts identically to B_SQUARE_CAP, except applied to joins.

SetLineMode() sets the line and join modes and the miter limit while LineCapMode(), LineJoinMode(), and LineMiterLimit() return them. The line mode affects all of the Stroke... methods except for Arc, Ellipse, and RoundRect.

SetLowColor() see SetHighColor()

SetOrigin(), Origin()

      void SetOrigin(BPoint pt)
      void SetOrigin(float x, float y)
      BPoint Origin(void) const

Sets and retrieves the local origin of the BView's coordinate system.

The actual origin used by the Application Server is the sum of the local origin (as set by this method) and the origins stored on the state stack (properly scaled).

SetPenSize(), PenSize()

      virtual void SetPenSize(float size)
      float PenSize(void) const

SetPenSize() sets the size of the BView's pen--the graphics parameter that determines the thickness of stroked lines--and PenSize() returns the current pen size. The pen size is stated in coordinate units, but is translated to a device-specific number of pixels for each output device.

The pen tip can be thought of as a brush that's centered on the line path and held perpendicular to it. If the brush is broader than one pixel, it paints roughly the same number of pixels on both sides of the path.

The default pen size is 1.0 coordinate unit. It can be set to any nonnegative value, including 0.0. If set to 0.0, the size is translated to one pixel for all devices. This guarantees that it will always draw the thinnest possible line no matter what the resolution of the device.

Thus, lines drawn with pen sizes of 1.0 and 0.0 will look alike on the screen (one pixel thick), but the line drawn with a pen size of 1.0 will be 1/72 of an inch thick when printed, however many printer pixels that takes, while the line drawn with a 0.0 pen size will be just one pixel thick.

The pen size is a parameter of the BView's graphics environment maintained by the Application Server and cached by the BView. If the BView isn't attached to a window, SetPenSize() records the size so that later, when the BView is added to a window and becomes known to the server, the cached value can automatically be established as the operable pen size for the BView. If the BView belongs to a window, this function changes both the server and the cache.

See also: "The Pen" in the "Drawing" section of this chapter, StrokeArc(), MovePenBy()

SetResizingMode(), ResizingMode()

      virtual void SetResizingMode(uint32 mode)
      uint32 ResizingMode(void) const

These functions set and return the BView's automatic resizing mode. The resizing mode is first set when the BView is constructed. The various possible modes are explained where the constructor is described.

See also: the BView constructor, SetFlags()

SetScale()

      void SetScale(float percent)

Sets the local scale of the BView's coordinate system. By default, each coordinate unit translates to one typographical point, about 1/72 of an inch. The percent argument scales the internal coordinate system of the view (not its frame rectangle) to a percentage of the default. For example, at a percent of 200.0, each coordinate unit will translate to 2 typographical points, 36 units per inch, and at a percent of 50.0, each unit will translate to a half point, 144 per inch.

The actual scale used by the Application Server is the product of the local clipping region (as set by this method) and the scaling factors stored on the state stack.

If the BView isn't attached to a window, this function will crash the application.

See also: the BPrintJob class

SetViewBitmap(), ClearViewBitmap()

      void SetViewBitmap(const BBitmap *bitmap, 
         BRect source, 
         BRect destination, 
         uint32 follow = B_FOLLOW_TOP | B_FOLLOW_LEFT, 
         uint32 options = B_TILE_BITMAP)
      void SetViewBitmap(const BBitmap *bitmap, 
         uint32 follow = B_FOLLOW_TOP | B_FOLLOW_LEFT, 
         uint32 options = B_TILE_BITMAP)
      void ClearViewBitmap(void)

SetViewBitmap() sets the background bitmap for the view. The view bitmap is a background image for the view; all drawing in the view occurs over this bitmap. The background color is used to fill in the visible regions not covered by the background bitmap.

The background bitmap is passed in bitmap. The caller is subsequently free to delete the bitmap; the Application Server will then keep the bitmap in memory as long as it is the background view.

If a source rectangle is given, only that part of the bitmap is used. Otherwise, the entire bitmap is used as the background bitmap. The destination bitmap, if given, specifies the placement of the bitmap in the view. It need not be the same size as the source rectangle; scaling is performed automatically by the application server. If no destination is given, the image will be placed, unscaled, at the upper left corner of the view.

follow determines the behavior of destination as the view is resized; see the BView constructor for specifics. options specifies additional view options. Currently, only one option, B_TILE_BITMAP, is defined. If set, the view bitmap is tiled across the view.

ClearViewBitmap() clears the background bitmap for the view.

See also: "The Background Bitmap" in the "Drawing" section of this chapter, SetViewColor()

SetViewColor(), ViewColor()

      virtual void SetViewColor(rgb_color color)
      inline void SetViewColor(uchar red, uchar green, uchar blue, uchar alpha = 0)
      rgb_color ViewColor(void) const

These functions set and return the background color that's shown in all areas of the view rectangle that the BView doesn't cover with its own drawing. When the clipping region is erased prior to an update, it's erased to the view color. When a view is resized to expose new areas, the new areas are first displayed in the view color. The default color is white, which matches the background color of the window's content area.

If you know that a BView will cover every pixel in the clipping region when it draws, you may want to avoid having the region erased to a color that will immediately be obliterated. If you set the view color to B_TRANSPARENT_32_BIT, the Application Server will not draw its background color before updates nor fill new areas with the background color. (Note that, despite the name, this doesn't make the view transparent--you can't see through it to what the view behind it would draw in that region.)

If you set the view color to anything but white (or B_TRANSPARENT_32_BIT), you must also set the B_WILL_DRAW flag. The flag informs the Application Server that there are specific drawing operations associated with the view. If it isn't set, the BView won't get update messages. Even if the BView does no other drawing--for example, if it doesn't implement a Draw() function--it must be updated to draw a background color other than white.

The version of SetViewColor() that takes separate arguments for the red, blue, and green color components works by creating an rgb_color data structure and passing it to the corresponding function that's declared virtual. Therefore, overriding only the virtual version will affect how both versions work. (However, due to the peculiarities of C++, overriding any version of an overloaded function hides all versions of the function. For continued access to the nonvirtual version without explicitly specifying the "BView::" prefix, simply copy the inline code from interface/View.h into the derived class.)

The alpha color component is currently ignored.

It's best to set the view color before the window is shown on-screen.

ViewColor() returns the current background color, which, because of internal caching mechanisms, doesn't normally entail contacting the Application Server.

See also: "The View Color" in the "Drawing" section of this chapter, SetHighColor(), SetViewBitmap()

Show() see Hide()

StringWidth(), GetStringWidths()

      float StringWidth(const char *string) const
      float StringWidth(const char *string, int32 length) const
      void GetStringWidths(char *stringArray[], int32 lengthArray[], int32 numStrings, float widthArray[]) const

These functions measure how much room is required to draw a string, or a group of strings, in the BView's current font. They're equivalent to the identically named set of functions defined in the BFont class, except that they assume the BView's font. For example, this line of code

   float width;
   width = myView->StringWidth("Be"B_UTF8_REGISTERED);

produces the same result as:

   float width;
   BFont font;
   myView->GetFont(&font);
   width = font.StringWidth("Be"B_UTF8_REGISTERED);

See the BFont class for details.

See also: BFont::StringWidth(), BFont::GetEscapements()

StrokeArc(), FillArc()

      void StrokeArc(BRect rect, float angle, float span, pattern aPattern = B_SOLID_HIGH)
      void StrokeArc(BPoint center, float xRadius, float yRadius, float angle, float span, pattern aPattern = B_SOLID_HIGH)
      void FillArc(BRect rect, float angle, float span, 
         pattern aPattern = B_SOLID_HIGH)
      void FillArc(BPoint center, float xRadius, float yRadius, float angle, float span, pattern aPattern = B_SOLID_HIGH)

These functions draw an arc, a portion of an ellipse. StrokeArc() strokes a line along the path of the arc. FillArc() fills the wedge defined by straight lines stretching from the center of the ellipse of which the arc is a part to the end points of the arc itself. For example:

The arc is a section of the ellipse inscribed in rect--or the ellipse located at center, where the horizontal distance from the center to the edge of the ellipse is measured by xRadius and the vertical distance from the center to the edge is measured by yRadius.

The arc starts at angle and stretches along the ellipse for span degrees, where angular coordinates are measured counterclockwise with 0 ° on the right, as shown below:

For example, if angle is 180.0 ° and span is 90.0 °, the arc would be the lower left quarter of the ellipse. The same arc would be drawn if angle were 270.0 ° and span were -90.0 °.

Currently, angle and span measurements in fractions of a degree are not supported.

The width of the line drawn by StrokeArc() is determined by the current pen size. Both functions draw using aPattern--or, if no pattern is specified, using the current high color. Neither function alters the current pen position.

See also: StrokeEllipse()

StrokeBezier(), FillBezier()

      void StrokeBezier(BPoint *controlPoints, pattern aPattern = B_SOLID_HIGH)
      void FillBezier(BPoint *controlPoints, pattern aPattern = B_SOLID_HIGH)

These functions draw a third degree Bezier curve. StrokeBezier() strokes a line along the path of the curve; the width of the line is determined by the current pen size. FillBezier() fills in the region defined by the path of the curve and the line joining the two endpoints.

controlPoints points to an array of the four points for the curve. Both functions draw using the pattern specified by aPattern--or, if no pattern is specified, in the current high color. Neither function alters the current pen position.

See also: SetPenSize(), StrokeRoundRect()

StrokeEllipse(), FillEllipse()

      void StrokeEllipse(BRect rect, pattern aPattern = B_SOLID_HIGH)
      void StrokeEllipse(BPoint center, float xRadius, float yRadius, pattern aPattern = B_SOLID_HIGH)
      void FillEllipse(BRect rect, pattern aPattern = B_SOLID_HIGH)
      void FillEllipse(BPoint center, float xRadius, float yRadius, pattern aPattern = B_SOLID_HIGH)

These functions draw an ellipse. StrokeEllipse() strokes a line around the perimeter of the ellipse and FillEllipse() fills the area the ellipse encloses.

The ellipse has its center at center. The horizontal distance from the center to the edge of the ellipse is measured by xRadius and the vertical distance from the center to the edge is measured by yRadius. If xRadius and yRadius are the same, the ellipse will be a circle.

Alternatively, the ellipse can be described as one that's inscribed in rect. If the rectangle is a square, the ellipse will be a circle.

The width of the line drawn by StrokeEllipse() is determined by the current pen size. Both functions draw using aPattern--or, if no pattern is specified, using the current high color. Neither function alters the current pen position.

See also: SetPenSize()

StrokeLine()

      void StrokeLine(BPoint start, BPoint end, pattern aPattern = B_SOLID_HIGH)
      void StrokeLine(BPoint end, pattern aPattern = B_SOLID_HIGH)

Draws a straight line between the start and end points--or, if no starting point is given, between the current pen position and end point--and leaves the pen at the end point.

This function draws the line using the current pen size and the specified pattern. If no pattern is specified, the line is drawn in the current high color. The points are specified in the BView's coordinate system.

See also: SetPenSize(), BeginLineArray()

StrokePolygon(), FillPolygon()

      void StrokePolygon(BPolygon *polygon, 
         bool isClosed = true, pattern aPattern = B_SOLID_HIGH)
      void StrokePolygon(BPoint *pointList, int32 numPoints, 
         bool isClosed = true, pattern aPattern = B_SOLID_HIGH)
      void StrokePolygon(BPoint *pointList, int32 numPoints, BRect rect, bool isClosed = true, pattern aPattern = B_SOLID_HIGH)
      void FillPolygon(BPolygon *aPolygon, 
         pattern aPattern = B_SOLID_HIGH)
      void FillPolygon(BPoint *pointList, int32 numPoints, pattern aPattern = B_SOLID_HIGH)
      void FillPolygon(BPoint *pointList, int32 numPoints, BRect rect, pattern aPattern = B_SOLID_HIGH)

These functions draw a polygon with an arbitrary number of sides. StrokePolygon() strokes a line around the edge of the polygon using the current pen size. If a pointList is specified rather than a BPolygon object, this function strokes a line from point to point, connecting the first and last points if they aren't identical. However, if the isClosed flag is false, StrokePolygon() won't stroke the line connecting the first and last points that define the BPolygon (or the first and last points in the pointList). This leaves the polygon open--making it not appear to be a polygon at all, but rather a series of straight lines connected at their end points. If isClosed is true, as it is by default, the polygon will appear to be a polygon, a closed figure.

FillPolygon() is a simpler function; it fills in the entire area enclosed by the polygon.

Both functions must calculate the frame rectangle of a polygon constructed from a point list--that is, the smallest rectangle that contains all the points in the polygon. If you know what this rectangle is, you can make the function somewhat more efficient by passing it as the rect parameter.

Both functions draw using the specified pattern--or, if no pattern is specified, in the current high color. Neither function alters the current pen position.

See also: SetPenSize(), the BPolygon class

StrokeRect(), FillRect()

      void StrokeRect(BRect rect, pattern aPattern = B_SOLID_HIGH)
      void FillRect(BRect rect, pattern aPattern = B_SOLID_HIGH)

These functions draw a rectangle. StrokeRect() strokes a line around the edge of the rectangle; the width of the line is determined by the current pen size. FillRect() fills in the entire rectangle.

Both functions draw using the pattern specified by aPattern--or, if no pattern is specified, in the current high color. Neither function alters the current pen position.

See also: SetPenSize(), StrokeRoundRect()

StrokeRoundRect(), FillRoundRect()

      void StrokeRoundRect(BRect rect, float xRadius, float yRadius, pattern aPattern = B_SOLID_HIGH)
      void FillRoundRect(BRect rect, float xRadius, float yRadius, pattern aPattern = B_SOLID_HIGH)

These functions draw a rectangle with rounded corners. The corner arc is one-quarter of an ellipse, where the ellipse would have a horizontal radius equal to xRadius and a vertical radius equal to yRadius.

Except for the rounded corners of the rectangle, these functions work exactly like StrokeRect() and FillRect().

Both functions draw using the pattern specified by aPattern--or, if no pattern is specified, in the current high color. Neither function alters the current pen position.

See also: StrokeRect(), StrokeEllipse()

StrokeTriangle(), FillTriangle()

      void StrokeTriangle(BPoint firstPoint, BPoint secondPoint, BPoint thirdPoint, pattern aPattern = B_SOLID_HIGH)
      void StrokeTriangle(BPoint firstPoint, BPoint secondPoint, BPoint thirdPoint, BRect rect, pattern aPattern = B_SOLID_HIGH)
      void FillTriangle(BPoint firstPoint, BPoint secondPoint, BPoint thirdPoint, pattern aPattern = B_SOLID_HIGH)
      void FillTriangle(BPoint firstPoint, BPoint secondPoint, BPoint thirdPoint, BRect rect, pattern aPattern = B_SOLID_HIGH)

These functions draw a triangle, a three-sided polygon. StrokeTriangle() strokes a line the width of the current pen size from the first point to the second, from the second point to the third, then back to the first point. FillTriangle() fills in the area that the three points enclose.

Each function must calculate the smallest rectangle that contains the triangle. If you know what this rectangle is, you can make the function marginally more efficient by passing it as the rect parameter.

Both functions do their drawing using the pattern specified by aPattern--or, if no pattern is specified, in the current high color. Neither function alters the current pen position.

See also: SetPenSize()

Sync() see Flush()

TargetedByScrollView()

      virtual void TargetedByScrollView(BScrollView *scroller)

Implemented by derived classes to respond to a notification that the BView has become the target of the scroller BScrollView object. This function is called when the BScrollView sets its target, which it does on construction. The target is the object whose contents will be scrolled.

BView's implementation of this function is empty.

See also: the BScrollView class

Window()

      BWindow *Window(void) const

Returns the BWindow to which the BView belongs, or NULL if the BView isn't attached to a window. This function returns the same object that Looper() (inherited from the BHandler class) does--except that Window() returns it more specifically as a pointer to a BWindow and Looper() returns it more generally as a pointer to a BLooper.

See also: BHandler::Looper() in the Application Kit, AddChild(), BWindow::AddChild(), AttachedToWindow()

WindowActivated()

      virtual void WindowActivated(bool active)

Implemented by derived classes to take whatever steps are necessary when the BView's window becomes the active window, or when the window gives up that status. If active is true, the window has become active. If active is false, it no longer is the active window.

All objects in the view hierarchy receive WindowActivated() notifications when the status of the window changes.

BView's version of this function is empty.

See also: BWindow::WindowActivated()

Scripting Support

The BView class implements the two suites "suite/vnd.Be-view" and "suite/vnd.Be-container-view." "suite/vnd.Be-container-view" is only implemented if the view contains a shelf; the suite consists of the single property "Shelf."

The Frame Property

Message Specifiers Meaning
B_GET_PROPERTY B_DIRECT_SPECIFIER Returns the view's frame rectangle.
B_SET_PROPERTY B_DIRECT_SPECIFIER Sets the view's frame rectangle.

Message	Specifiers	Meaning
`B_GET_PROPERTY`	`B_DIRECT_SPECIFIER`	Returns the view's frame rectangle.
`B_SET_PROPERTY`	`B_DIRECT_SPECIFIER`	Sets the view's frame rectangle.

The "Frame" property represents the frame rectangle of the view. The frame is passed as a BRect (B_RECT_TYPE).

The Hidden Property

Message Specifiers Meaning
B_GET_PROPERTY B_DIRECT_SPECIFIER Returns true if the view is hidden; false otherwise.
B_SET_PROPERTY B_DIRECT_SPECIFIER Hides or shows the view.

Message	Specifiers	Meaning
`B_GET_PROPERTY`	`B_DIRECT_SPECIFIER`	Returns `true` if the view is hidden; `false` otherwise.
`B_SET_PROPERTY`	`B_DIRECT_SPECIFIER`	Hides or shows the view.

The "Hidden" property determines the visibility of the view. The messages are equivalent to manipulating the view with the IsHidden(), Hide(), and Show(). Note that this differs slightly from BWindow's "Hidden" property, where multiple hides or shows are not nested.

The "Hidden" property determines the visibility of the view. The messages are equivalent to manipulating the view with the IsHidden(), Hide(), and Show() methods with one caveat: nested Hide() and Show() calls are disabled so that multiple scripting Hide commands may be undone with a single Show.

The Shelf Property

Message Specifiers Meaning
any B_DIRECT_SPECIFIER Directs the scripting message to the shelf.

Message	Specifiers	Meaning
any	`B_DIRECT_SPECIFIER`	Directs the scripting message to the shelf.

The "Shelf" property pops the current specifier off the specifier stack and then passes the scripting message to the shelf. If no shelf is present, an error is returned.

The View Property

Message Specifiers Meaning
B_COUNT_PROPERTIES B_DIRECT_SPECIFIER Returns the number of of child views.
any B_INDEX_SPECIFIER,B_REVERSE_INDEX_SPECIFIER,
B_NAME_SPECIFIER Directs the scripting message to the specified view.

Message	Specifiers	Meaning
`B_COUNT_PROPERTIES`	`B_DIRECT_SPECIFIER`	Returns the number of of child views.
any	`B_INDEX_SPECIFIER,`B_REVERSE_INDEX_SPECIFIER, B_NAME_SPECIFIER	Directs the scripting message to the specified view.

The "View" property represents the child views of the current view. For all messages except B_COUNT_PROPERTIES, the current specifier is popped off the specifier stack before the scripting message is passed to the target view. Views can be specified either by index (as found by ChildAt()), or name (as found by FindView()).

Archived Fields

The Archive() function adds the following fields to its BMessage argument:

Field Type code Meaning
"_frame" B_RECT_TYPE The view's frame rectangle.
"_resize_mode" B_INT32_TYPE Resizing mode.
"_flags" B_INT32_TYPE View flags.
"_fname" (array) B_STRING_TYPE Font family and style.
"_fflt" (array) B_FLOAT_TYPE Font size, shear, and rotation (-1 if default).
"_color" (array) B_INT32_TYPE High, low, and view colors.
"_psize" B_FLOAT_TYPE Pen size.
"_dmod" B_INT32_TYPE Drawing mode.
"_views" B_MESSAGE_TYPE Archived child views (deep copy only).

Field	Type code	Meaning
"_frame"	`B_RECT_TYPE`	The view's frame rectangle.
"_resize_mode"	`B_INT32_TYPE`	Resizing mode.
"_flags"	`B_INT32_TYPE`	View flags.
"_fname" (array)	`B_STRING_TYPE`	Font family and style.
"_fflt" (array)	`B_FLOAT_TYPE`	Font size, shear, and rotation (-1 if default).
"_color" (array)	`B_INT32_TYPE`	High, low, and view colors.
"_psize"	`B_FLOAT_TYPE`	Pen size.
"_dmod"	`B_INT32_TYPE`	Drawing mode.
"_views"	`B_MESSAGE_TYPE`	Archived child views (deep copy only).

Some of these fields may not be present if the setting they represent isn't used, or is the default value. For example, if the font's family and style were not changed, the "_fname" field won't be found in the archive.

The Be Book, in lovely HTML, for BeOS Release 4.

Last modified December 23, 1998.

`B_TRACK_WHOLE_RECT`	The whole rectangle moves with the cursor. Its position changes, but its size remains fixed.
`B_TRACK_RECT_CORNER`	The left top corner of the rectangle remains fixed within the view while its right and bottom edges move with the cursor.

`B_BACKSPACE`	`B_LEFT_ARROW`	`B_INSERT`
`B_ENTER`	`B_RIGHT_ARROW`	`B_DELETE`
`B_RETURN`	`B_UP_ARROW`	`B_HOME`
`B_SPACE`	`B_DOWN_ARROW`	`B_END`
`B_TAB`		`B_PAGE_UP`
`B_ESCAPE`	`B_FUNCTION_KEY`	`B_PAGE_DOWN`

`B_F1_KEY`	`B_F6_KEY`	`B_F11_KEY`
`B_F2_KEY`	`B_F7_KEY`	`B_F12_KEY`
`B_F3_KEY`	`B_F8_KEY`	`B_PRINT_KEY` (Print Screen)
`B_F4_KEY`	`B_F9_KEY`	`B_SCROLL_KEY` (Scroll Lock)
`B_F5_KEY`	`B_F10_KEY`	`B_PAUSE_KEY`

`B_OP_COPY`	`B_OP_MIN`
`B_OP_OVER`	`B_OP_MAX`
`B_OP_ERASE`	`B_OP_ADD`
`B_OP_INVERT`	`B_OP_SUBTRACT`
`B_OP_SELECT`	`B_OP_BLEND`

`B_FONT_FAMILY_AND_STYLE`	`B_FONT_SPACING`
`B_FONT_SIZE`	`B_FONT_ENCODING`
`B_FONT_SHEAR`	`B_FONT_FACE`
`B_FONT_ROTATION`	`B_FONT_FLAGS`

`B_ROUND_CAP`	A semicircle is drawn around the endpoint. Its diameter is equal to the width of the line.
`B_BUTT_CAP`	The line is squared off and does not extend beyond the endpoint.
`B_SQUARE_CAP`	The line is squared off, extending past the endpoint for a distance equal to half the width of the line.

`B_ROUND_JOIN`	Acts identically to `B_ROUND_CAP`, except applied to joins.
`B_MITER_JOIN`	The lines are extended until they touch. If they meet at an angle greater than 2arcsin(1/miterLimit*), a bevel join is used instead.
`B_BEVEL_JOIN`	Butt end caps are used at the common endpoint and the empty area between the caps is filled with a triangle.
`B_BUTT_JOIN`	Acts identically to `B_BUTT_CAP`, except applied to joins.
`B_SQUARE_JOIN`	Acts identically to `B_SQUARE_CAP`, except applied to joins.