Ports
Volume Number: | | 1
|
Issue Number: | | 4
|
Column Tag: | | C Workshop
|
Ports
By Chris Derossi
When one thinks of the Macintosh, one of the first things to come to mind is the Macs frequent use of windows for multiple, independent areas or objects. Usually, one window is related to one task or concept, and tasks that have very different orientations are in separate windows.
There are several benefits for grouping items or tasks into separate windows. The most important is user interfacing. By using windows, the user can conceptually group similar or common things. This makes the use of sotware more intuitive because the user is shielded from modes in the software. In other words, instead of explicitly needing to have the computer switch to a different mode, all that is needed is to begin operations in a different window.
Obviously, the more intuitive for the user that your software is, the better. Indeed, one of the primary concepts of the Macintosh is that of using tools like icons, the mouse, and windows to decrease the need for special knowledge or special education. In this article, we will take a look at the concept behind windows, that of ports. (In a later article, we will examine windows more closely.)
In order to facilitate the creation of independent windows, the implementation of ports or graph ports provides for completely separate drawing envir- onments. That is, each port is an entire drawing environment, and operations with one port are not related to another.
At any given time, after QuickDraw is initialized, a single port is always active, and is considered the current graph port. QuickDraw operations always occur within the current graph port.
With each graph port are associated several characteristics and parameters. These include location of the ports bit map, size, clipping region, visible region, background color, and several text mode values. The ports bit map refers to an area of memory where the bit image of any drawing is put. Usually, this coincides with an area of the screen ram, so that drawing is displayed on the screen. However, this is not a requirement. The ports bit map can be any area of memory on or off the screen. For example, a port may use an off-screen bit map to prepare an image, then move that image into the screen bit map.
The size of the port is determined by an enclosing rectangle. The rectangle encloses part of the bit map and indicates size and position within the bit map. For ports that are visible, the rectangle encloses an area of the screen.
Clipping region and visible region both restrict the available drawing area within a graph port. While the port itself is rectangular, the clipping and visible areas are regions, providing total flexibility with shape and configuration. The visible region acts like a clipping region, but is separate from it mainly for use by the window manager. (When one window overlaps another, the visible region of the bottom window is set to equal the part of the port that still shows.) There are two regions so that the user may have a region for clipping that is unrelated to the the region needed by the window manager for handling overlap.
Each port is its own drawing environment, and as such has its own drawing characteristics such as pen size, pattern, position, etc. Also, each port has its own text drawing characteristics that are used by the font manager. This way, you need only choose which port in which to draw, and the font, style, and text size associated with that port are automatically used.
In addition to having its own locations, size, and modes, each graph port may also have its own coordinate system. That is, regardless of where on the screen the port is displayed, the upper left corner of the port may be any coordinate. Usually, of course, the upper left coordinate is assigned the value 0,0. This is useful because neither the user nor the programmer need be concerned with the actual position of any port; drawing may be made as if the port were the entire universe. (You might begin to see why moving windows around is so easy; the coordinates inside the window remain constant.)
Since the available coordinate plane is much larger than any visible graph port, this feature allows ports to look at different areas of the coordinate plane. For example, a program may draw on the entire plane, and different graph ports may be created as neccessary to view separate parts of the entire area. This can be done without the need for conversions either in the drawing or the displaying; QuickDraw handles all conversions.
There are two important programming practices to develop with regards to ports. The first has already been mentioned; it is a good idea to group similar things, and separate dissimilar things. The second idea is that of port independence. Programs should be written to not rely on any port size or location. (When changing the size of a window, for instance, the program should NOT need a certain size window to perform.)
The sample program this month illustrates the second concept, as the first one is much easier to practice. In this program, the user creates television sets on which drawing is displayed. While drawing is occurring, the user may create more TVs, or change which one is active. All this is done by clicking the mouse. If the mouse is clicked in a TV, it becomes the active one, if not, a new TV is created.
Each TV, of course, is a graph port. A port is created, and then in that port a television is drawn. Then, the clipping region of that port is restricted to the picture tube part of the picture.
Drawing continues, regardless of which port is active, or where it is. The placement of the port, and the selection of the current port is entirely up to the user. Notice, that the main drawing procedure has absolutely no idea which port is active.
In order to avoid running out of memory, the number of possible ports is restricted. You may vary the constant MaxPorts as you like. An array is set up called Screens which contains pointers to the graph ports. QuickDraw procedures and functions that act upon ports take a pointer to the port as the parameter, not the port itself. The port is left to be dynamically allocated for flexibility. The type GrafPtr is the predefined pointer to a graph port.
The variable ScrnCnt keeps track of how many ports we have created. Since we want to draw only on our TV screens, drawing does not occur if the user has not created any ports yet. When the user clicks the mouse, the program checks to see if the position of the mouse is inside any graph port. The point must first be brought to global coordinates, which makes it independent of any port. Then, one by one, the point is referenced to each graph port and checked against that ports rectangle. If the point is inside the rectangle, that port is selected as the current one.
When the point lies outside all ports, a new port is created. (Unless all possible ports have been created already; then the program terminates.) The port is initialized and its characteristics are set. Then, the TV is drawn in the port, and the clipping region set. The new port is also set as the current port. Notice that the program explicitly sets the ports origin. Try playing with the origin values.
When the program ends, it closes each port one by one and frees the memory used by the ports. Then, it creates another temporary port that covers the entire screen. (The default values for the size and location of the port coincide with the screen.) This port is then filled with gray to eliminate the other graphics. Finally, this temporary port is also closed and its memory freed.
Although this is a simple example, it demonstrates a very powerful principle; that of independence. The more independent you programs are, the more flexible they are, and the more control the user may have. This avoids placing pointless limitations on your programs and their users. In addition, programming of this nature helps keep programs functional when the environment changes. (Like moving from a Mac to the Lisa with MacWorks, or future, improved machines.)
This concludes our introduction to the basics of QuickDraw. With this foundation, we can move on into the rest of the Macintosh with a strong background. You might want to try and improve on this sample program as an excersize to increase your skill. For example, try to allow the user to select the size as well as the location of the port. Or, try to prevent the ports from overlapping. If youre really enthusiastic, have all the TV sets going at once, and let the user change channels on each of them, causing different graphics to appear on each channel.
In the next issue, well extend the discussion of ports to windows. Well find out what the concept of windows adds to the already powerful idea of ports, and find out what features are provided to the user through the use of windows. Ciao.
program Ports_Demo;
{ Ports_Demo - for MacTutor}
{ -- by Chris Derossi}
uses
QuickDraw2; {QuickDraw2 contains the stuff for regions.}
const
MaxScreens = 8;
type
TVList = array[1..MaxScreens] of GrafPtr;
var
Screens : TVList;
ScrnCnt : INTEGER;
done : BOOLEAN;
procedure SetUp;
begin
HideAll;
ScrnCnt := 0;
done := FALSE;
end;
procedure ShutDown;
var
TempPort : GrafPtr;
begin
while ScrnCnt > 0 do
begin
ClosePort(Screens[ScrnCnt]);
Dispose(Screens[ScrnCnt]);
ScrnCnt := ScrnCnt - 1;
end;
{ Create a new port and fill the screen with gray. }
NEW(TempPort);
OpenPort(TempPort);
BackPat(Gray);
EraseRect(TempPort^.PortRect);
ClosePort(TempPort);
Dispose(TempPort);
end;
function MakeTV (Left, Top : INTEGER) : GrafPtr;
var
TempPtr : GrafPtr;
WorkRect : Rect;
begin
{ Create a new port and set its characteristics. }
NEW(TempPtr);
OpenPort(TempPtr);
PortSize(100, 80);
MovePortTo(Left, Top);
SetOrigin(0, 0);
{ Clear the port to white and draw our television }
BackPat(White);
EraseRect(TempPtr^.PortRect);
FrameRect(TempPtr^.PortRect);
MoveTo(80, 0);
LineTo(80, 80);
SetRect(WorkRect, 85, 10, 95, 20);
FrameOval(WorkRect);
SetRect(WorkRect, 85, 30, 95, 40);
FrameOval(WorkRect);
{ Restrict the clipping region to the TV screen }
WorkRect := TempPtr^.PortRect;
WorkRect.right := WorkRect.Right - 20;
InsetRect(WorkRect, 1, 1);
ClipRect(WorkRect);
MakeTV := TempPtr;
end;
procedure SetTV;
{ If the mouse is in one of our ports, select that port. If all}
{possible ports are created and the mouse is not in any port,}
{then were done. Otherwise, create a new port.}
var
MousePt, Pnt1 : Point;
I, X, Y : INTEGER;
TempPtr : GrafPtr;
begin
TempPtr := nil;
GetMouse(X, Y); { Current ports local coordinates }
repeat { nothing }
until not Button;
{ Convert to a point, the to global coordinates }
MousePt.h := X;
MousePt.v := Y;
LocalToGlobal(MousePt);
Pnt1 := MousePt;
{ Scan the existing ports. }
if ScrnCnt > 0 then { we have some ports to scan }
for I := 1 to ScrnCnt do
begin
SetPort(Screens[I]);
MousePt := Pnt1;
GlobalToLocal(MousePt);
if PtInRect(MousePt, Screens[I]^.PortRect) then
TempPtr := Screens[I];
end; { for loop }
if TempPtr <> nil then { the mouse is in a port; set it }
SetPort(TempPtr)
else if ScrnCnt = MaxScreens then { all ports used }
done := TRUE
else
begin
ScrnCnt := ScrnCnt + 1;
Screens[ScrnCnt] := MakeTV(Pnt1.h, Pnt1.v);
end;
end;
procedure MainLoop;
var
Figure : INTEGER;
WorkRect : Rect;
begin
if Button then
SetTV;
if ScrnCnt > 0 then { do some drawing }
begin
if random mod 100 < 5 then { Clear the screen }
begin
SetRect(WorkRect, 0, 0, 90, 90);
EraseRect(WorkRect);
end;
{ Create a random rectangle for drawing }
WorkRect.top := random mod 80;
WorkRect.left := random mod 80;
WorkRect.right := WorkRect.left + (random mod 60);
WorkRect.bottom := WorkRect.top + (random mod 60);
Figure := random mod 6;
case Figure of
0 :
FrameRect(WorkRect);
1 :
FrameRoundRect(WorkRect, 18, 18);
2 :
FrameOval(WorkRect);
3 :
PaintRect(WorkRect);
4 :
PaintRoundRect(WorkRect, 18, 18);
5 :
PaintOval(WorkRect);
end;
end;
end;
begin { Ports_Demo }
SetUp;
while not done do
MainLoop;
ShutDown;
end.