Thinking about Bits

[variable]<<[number of places] 

int mult_by_pow_2(int number, int power) {     return number<Note that in this example, we're using integers, which are either 2 or 4 bytes, and that the operation gets applied to the entire sequence of 16 or 32 bits.

But what happens if we shift a number like 128 and we're only storing it in a single byte: 10000000? Well, 128 * 2 = 256, and we can't even store a number that big in a byte, so it shouldn't be surprising that the result is 00000000.

It shouldn't surprise you that there's a corresponding right-shift operator: >> (especially considering that I mentioned it earlier). Note that a bitwise right-shift will be the equivalent of integer division by 2.

Why is it integer division? Consider the number 5, in binary, 00000101. 5/2 is 2.5, but if you are performing integer division, 5/2 is 2. When you perform a right shift by one: (unsigned int)5>>1, you end up with 00000010, as the rightmost 1 gets shifted off the end; this is the representation of the number 2. Note that this only holds true for unsigned integers; otherwise, we are not guaranteed that the padding bits will be all 0s.

Generally, using the left and right shift operators will result in significantly faster code than calculating and then multiplying by a power of two. The shift operators will also be useful later when we look at how to manipulating individual bits.

For now, let's look at some of the other binary operators to see what they can do for us.
Bitwise AND
The bitwise AND operator is a single ampersand: &. A handy mnemonic is that the small version of the boolean AND, &&, works on smaller pieces (bits instead of bytes, chars, integers, etc). In essence, a binary AND simply takes the logical AND of the bits in each position of a number in binary form.

For instance, working with a byte (the char type):
01001000 &  10111000 =  -------- 00001000 
The most significant bit of the first number is 0, so we know the most significant bit of the result must be 0; in the second most significant bit, the bit of second number is zero, so we have the same result. The only time where both bits are 1, which is the only time the result will be 1, is the fifth bit from the left. Consequently,
72 & 184 = 8 
Bitwise OR
Bitwise OR works almost exactly the same way as bitwise AND. The only difference is that only one of the two bits needs to be a 1 for that position's bit in the result to be 1. (If both bits are a 1, the result will also have a 1 in that position.) The symbol is a pipe: |. Again, this is similar to boolean logical operator, which is ||.
01001000 |  10111000 =  -------- 11111000 
and consequently
72 | 184 = 248 
Let's take a look at an example of when you could use just these four operators to do something potentially useful. Let's say that you wanted to keep track of certain boolean attributes about something -- for instance, you might have eight cars (!) and want to keep track of which are in use. Let's assign each of the cars a number from 0 to 7.

Since we have eight items, all we really need is a single byte, and we'll use each of its eight bits to indicate whether or not a car is in use. To do this, we'll declare a char called in_use, and set it to zero. (We'll assume that none of the cars are initially "in use".)
char in_use = 0; 
Now, how can we check to make sure that a particular car is free before we try to use it? Well, we need to isolate the one bit that corresponds to that car. The strategy is simple: use bitwise operators to ensure every bit of the result is zero except, possibly, for the bit we want to extract.

Consider trying to extract the fifth bit from the right of a number: XX?XXXXX We want to know what the question mark is, and we aren't concerned about the Xs. We'd like to be sure that the X bits don't interfere with our result, so we probably need to use a bitwise AND of some kind to make sure they are all zeros. What about the question mark? If it's a 1, and we take the bitwise AND of XX?XXXXX and 00100000, then the result will be 00100000:
XX1XXXXX &  00100000 =  -------- 00100000 
Whereas, if it's a zero, then the result will be 00000000:
XX0XXXXX &  00100000 =  -------- 00000000 
So we get a non-zero number if, and only if, the bit we're interested in is a 1.

This procedure works for finding the bit in the nth position. The only thing left to do is to create a number with only the one bit in the correct position turned on. These are just powers of two, so one approach might be to do something like:
int is_in_use(int car_num) {     // pow returns an int, but in_use will also be promoted to an int     // so it doesn't have any effect; we can think of this as an operation     // between chars     return in_use & pow(2, car_num); } 
While this function works, it can be confusing. It obscures the fact that what we want to do is shift a bit over a certain number of places, so that we have a number like 00100000 -- a couple of zeros, a one, and some more zeros. (The one could also be first or last -- 10000000 or 00000001.)

We can use a bitwise leftshift to accomplish this, and it'll be much faster to boot. If we start with the number 1, we are guaranteed to have only a single bit, and we know it's to the far-right. We'll keep in mind that car 0 will have its data stored in the rightmost bit, and car 7 will be the leftmost.
int is_in_use(int car_num) {     return in_use & 1<Note that shifting by zero places is a legal operation -- we'll just get back the same number we started with.

All we can do right now is check whether a car is in use; we can't actually set the in-use bit for it. There are two cases to consider: indicating a car is in use, and removing a car from use. In one case, we need to turn a bit on, and in the other, turn a bit off.

Let's tackle the problem of turning the bit on. What does this suggest we should do? If we have a bit set to zero, the only way we know right now to set it to 1 is to do a bitwise OR. Conveniently, if we perform a bitwise OR with only a single bit set to 1 (the rest are 0), then we won't affect the rest of the number because anything ORed with zero remains the same (1 OR 0 is 1, and 0 OR 0 is 0).

Again we need to move a single bit into the correct position: void set_in_use(int car_num) { in_use = in_use | 1<The Bitwise ComplementThe bitwise complement operator, the tilde, ~, flips every bit. A useful way to remember this is that the tilde is sometimes called a twiddle, and the bitwise complement twiddles every bit: if you have a 1, it's a 0, and if you have a 0, it's a 1.

This turns out to be a great way of finding the largest possible value for an unsigned number:
unsigned int max = ~0; 
0, of course, is all 0s: 00000000 00000000. Once we twiddle 0, we get all 1s: 11111111 11111111. Since max is an unsigned int, we don't have to worry about sign bits or twos complement. We know that all 1s is the largest possible number.

Note that ~ and ! cannot be used interchangeably. When you take the logical NOT of a non-zero number, you get 0 (FALSE). However, when you twiddle a non-zero number, the only time you'll get 0 is when every bit is turned on. (This non-equivalence principle holds true for bitwise AND too, unless you know that you are using strictly the numbers 1 and 0. For bitwise OR, to be certain that it would be equivalent, you'd need to make sure that the underlying representation of 0 is all zeros to use it interchangeably. But don't do that! It'll make your code harder to understand.)

Now that we have a way of flipping bits, we can start thinking about how to turn off a single bit. We know that we want to leave other bits unaffected, but that if we have a 1 in the given position, we want it to be a 0. Take some time to think about how to do this before reading further.

We need to come up with a sequence of operations that leaves 1s and 0s in the non-target position unaffected; before, we used a bitwise OR, but we can also use a bitwise AND. 1 AND 1 is 1, and 0 AND 1 is 0. Now, to turn off a bit, we just need to AND it with 0: 1 AND 0 is 0. So if we want to indicate that car 2 is no longer in use, we want to take the bitwise AND of XXXXX1XX with 11111011.

How can we get that number? This is where the ability to take the complement of a number comes in handy: we already know how to turn a single bit on. If we turn one bit on and take the complement of the number, we get every bit on except that bit:
~(1<Now that we have this, we can just take the bitwise AND of this with the current field of cars, and the only bit we'll change is the one of the car_num we're interested in.
int set_unused(int car_num) {     in_use = in_use & ~(1<You might be thinking to yourself, but this is kind of clunky. We actually need to know whether a car is in use or not (if the bit is on or off) before we can know which function to call. While this isn't necessarily a bad thing, it means that we do need to know a little bit about what's going on. There is an easier way, but first we need the last bitwise operator: exclusive-or.
Bitwise Exclusive-Or (XOR)
There is no boolean operator counterpart to bitwise exclusive-or, but there is a simple explanation. The exclusive-or operation takes two inputs and returns a 1 if either one or the other of the inputs is a 1, but not if both are. That is, if both inputs are 1 or both inputs are 0, it returns 0. Bitwise exclusive-or, with the operator of a carrot, ^, performs the exclusive-or operation on each pair of bits. Exclusive-or is commonly abbreviated XOR.

For instance, if you have two numbers represented in binary as 10101010 and 01110010 then taking the bitwise XOR results in 11011000. It's easier to see this if the bits are lined up correctly:
01110010 ^ 10101010  -------- 11011000 
You can think of XOR in the following way: you have some bit, either 1 or 0, that we'll call A. When you take A XOR 0, then you always get A back: if A is 1, you get 1, and if A is 0, you get 0. On the other hand, when you take A XOR 1, you flip A. If A is 0, you get 1; if A is 1, you get 0.

So you can think of the XOR operation as a sort of selective twiddle: if you apply XOR to two numbers, one of which is all 1s, you get the equivalent of a twiddle.

Additionally, if you apply the XOR operation twice -- say you have a bit, A, and another bit B, and you set C equal to A XOR B, and then take C XOR B: you get A XOR B XOR B, which essentially either flips every bit of A twice, or never flips the bit, so you just get back A. (You can also think of B XOR B as cancelling out.) As an exercise, can you think of a way to use this to exchange two integer variables without a temporary variable? (Once you've figured it out, check the solution.)

How does that help us? Well, remember the first principle: XORing a bit with 0 results in the same bit. So what we'd really like to be able to do is just call one function that flips the bit of the car we're interested in -- it doesn't matter if it's being turned on or turned off -- and leaves the rest of the bits unchanged.

This sounds an awful lot like the what we've done in the past; in fact, we only need to make one change to our function to turn a bit on. Instead of using a bitwise OR, we use a bitwise XOR. This leaves everything unchanged, but flips the bit instead of always turning it on:
void flip_use_state(int car_num) {     in_use = in_use ^ 1<
When should you use bitwise operators?
Bitwise operators are good for saving space -- but many times, space is hardly an issue. And one problem with working at the level of the individual bits is that if you decide you need more space or want to save some time -- for instance, if we needed to store information about 9 cars instead of 8 -- then you might have to redesign large portions of your program. On the other hand, sometimes you can use bitwise operators to cleverly remove dependencies, such as by using ~0 to find the largest possible integer. And bit shifting to multiply by two is a fairly common operation, so it doesn't affect readability in the way that advanced use of bit manipulation can in some cases (for instance, using XOR to switch the values stored in two variables). 

Introduction to Windows Programming and OpenGL:

#include  int APIENTRY WinMain(HINSTANCE hInstance,                      HINSTANCE hPrevInstance,                      LPSTR     lpCmdLine,                      int       nCmdShow) {     MessageBox(NULL, "\tHello World!", "My first windows app", NULL);  return 0; }

Every Windows programming needs a main entry point. That being said, you may be wondering what this entry point is. The main entry point for any Windows program is called WinMain. The function prototype for WinMain is a little confusing at first, but as we continue to work with it you'll notice it becomes much easier to understand. Well. we've told you what it is; now were going to show you! Here's the prototype for WinMain:

int WINAPI WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, LPSTR lpCmdLine, int nShowCmd);

hInstance - is a handle to your applications instance, where an instance  can be considered to be a single run of your application. The instance  is used by windows as a reference to your application for event handling, message processing, and various other duties. hPrevInstance - is always NULL.  lpCmdLine - is a pointer string that is used to hold any command-line arguments that may have been specified when the application began.   For example, if the user opened the Run application and typed myapp.exe myparameter 1, then lpCmdLine would be myparameter 1. nShowCMD - is the parameter that determines how your application's window will be displayed once it begins executing. 

LRESULT CALLBACK WndProc(HWND hwnd, UINT message, WPARAM wParam, LPARAM lParam); 

hwnd - Only important if you have several windows of the same class open at one time. This is used to determine which window hwnd pointed to before  deciding on an action. message - The actual message identifier that WndProc will be handling. wParam and lParam - Extensions of the message parameter. Used to give  more information and point to specifics that message cannot on its own. 

First Windows Application:

/* Trim fat from windows*/ #define WIN32_LEAN_AND_MEAN  #pragma comment(linker, "/subsystem:windows") /* Pre-processor directives*/ #include "stdafx.h" #include  /* Windows Procedure Event Handler*/ LRESULT CALLBACK WndProc(HWND hwnd, UINT message, WPARAM wParam, LPARAM lParam) {  PAINTSTRUCT paintStruct;  /* Device Context*/  HDC hDC;   /* Text for display*/  char string[] = "Hello, World!";   /* Switch message, condition that is met will execute*/  switch(message)  {   /* Window is being created*/   case WM_CREATE:     return 0;    break;   /* Window is closing*/   case WM_CLOSE:     PostQuitMessage(0);    return 0;    break;   /* Window needs update*/   case WM_PAINT:     hDC = BeginPaint(hwnd,&paintStruct);    /* Set txt color to blue*/    SetTextColor(hDC, COLORREF(0x00FF0000));    /* Display text in middle of window*/    TextOut(hDC,150,150,string,sizeof(string)-1);    EndPaint(hwnd, &paintStruct);    return 0;    break;   default:    break;  }  return (DefWindowProc(hwnd,message,wParam,lParam)); } /* Main function*/ int APIENTRY WinMain(HINSTANCE hInstance,                      HINSTANCE hPrevInstance,                      LPSTR     lpCmdLine,                      int       nCmdShow) {  WNDCLASSEX  windowClass;  //window class  HWND  hwnd;    //window handle  MSG   msg;    //message  bool  done;    //flag saying when app is complete  /* Fill out the window class structure*/  windowClass.cbSize = sizeof(WNDCLASSEX);  windowClass.style = CS_HREDRAW | CS_VREDRAW;  windowClass.lpfnWndProc = WndProc;  windowClass.cbClsExtra = 0;  windowClass.cbWndExtra = 0;  windowClass.hInstance = hInstance;  windowClass.hIcon = LoadIcon(NULL, IDI_APPLICATION);  windowClass.hCursor = LoadCursor(NULL, IDC_ARROW);  windowClass.hbrBackground = (HBRUSH)GetStockObject(WHITE_BRUSH);  windowClass.lpszMenuName = NULL;  windowClass.lpszClassName = "MyClass";  windowClass.hIconSm = LoadIcon(NULL, IDI_WINLOGO);  /* Register window class*/  if (!RegisterClassEx(&windowClass))  {   return 0;  }  /* Class registerd, so now create window*/  hwnd = CreateWindowEx(NULL,  //extended style   "MyClass",   //class name   "A Real Win App",  //app name   WS_OVERLAPPEDWINDOW |  //window style   WS_VISIBLE |   WS_SYSMENU,   100,100,   //x/y coords   400,400,   //width,height   NULL,    //handle to parent   NULL,    //handle to menu   hInstance,   //application instance   NULL);    //no extra parameter's  /* Check if window creation failed*/  if (!hwnd)   return 0;  done = false; //initialize loop condition variable  /* main message loop*/  while(!done)  {   PeekMessage(&msg,NULL,NULL,NULL,PM_REMOVE);   if (msg.message == WM_QUIT) //check for a quit message   {    done = true; //if found, quit app   }   else   {    /* Translate and dispatch to event queue*/    TranslateMessage(&msg);     DispatchMessage(&msg);   }  }  return msg.wParam; } 

HDC hDC;    //device context Char string[] = "Hello World!"; //display text 

case WM_PAINT:     hDC = BeginPaint(hwnd,&paintStruct);    /* Set txt color to blue*/    SetTextColor(hDC, COLORREF(0x00FF0000));    /* Display text in middle of window*/    TextOut(hDC,150,150,string,sizeof(string)-1);    EndPaint(hwnd, &paintStruct);    return 0;    break; 

1. Window-class setup 2. Window-class registration 3. Window Creation 4. Message loop with event handler 

WGL and the Wiggle Functions in C++:

• wglCreateContext();
• wglDeleteContext();
• wglMakeCurrent();

 HGLRC wglCreateContext(HDC hDC); 

 BOOL  wglDeleteContext(HGLRC hRC); 

 BOOL  wglMakeCurrent(HDC hDC, HGLRC hRC); 

LRESULT CALLBACK WndProc (HWND hwnd, UNIT message, WPARAM wParam, LPARAM lParam) {  static HGLRC hRC; //rendering context  static HDC hDC; //device context  switch (message)  {  case WM_CREATE:  hDC = GetDC(hwnd); //get the device context for window  hRC = wglCreateContext(hDC); //create rendering context  wglMakeCurrent(hDC,hRC); //make rendering context current  break;  case WM_DESTROY:  wglMakeCurrent(hDC,NULL); //deselect rendering context  wglDeleteContext(hRC);  //delete rendering context  PostQuitMessage(0);  //send wm_quit  break;  } }