IC210: Keep it classy (or structy?) — a little fun with curses

Part 1: ncurses basics (lab11v1.cpp)

The ncurses library views your terminal window as a big 2D table of single-character positions. So think of coordinates as (row,col) rather than (x,y). Thinking of it this way, (0,0) is the upper left corner of the terminal window. To use ncurses, you need to do a #include <curses.h> in your .cpp files, and you need to compile with the -l ncurses compiler flag, like g++ lab11v1.cpp -l ncurses.

Create and compile lab11v1.cpp under your account. This source code demonstrates some simple aspects of ncurses. Here's a few comments:

Initializing Ncurses	Exiting Ncurses	Drawing something
WINDOW *W = initscr(); cbreak(); noecho(); nodelay(W,true); curs_set(0);	// Loop forever until user enters 'q' char c; while((c = wgetch(W)) && c != 'q'); // Clean up after ncurses endwin();	wmove(W,0,0); waddch(W,'1'); wrefresh(W); usleep(800000);
This code should appear at the point in your main when you're ready to go into "ncurses mode". When this is executed, the screen goes blank and you are ready to write to arbitrary locations in the terminal window. The variable `W` is important. It's a pointer to a struct object that the ncurses library defines. You'll need to pass `W` to many ncurses functions.	Ncurses (at least as we've initialized it) changes the model for how we read data. You fetch the next character with a call to `wgetch(W)`. The funny thing is that this is "non-blocking I/O", which means that you don't hang around and wait for input to come. If you call `wgetch(W)` and nobody presses a key, wgetch simply returns immediate with an error code. Contrast this with `cin >> c` or `cin.get()`. Thus, this loop just spins waiting for wgetch(W) to report that the user has pressed q. The `endwin()` function call is required in order for ncurses to clean up after itself and restore the terminal window.	The model for writing to the screen in ncurses is this: move the cursor to a position the terminal window grid int wmove(Window* W, int row, int col); write a character at the current cursor position (which has the side effect of incrementing the cursor's column coordinate) int waddch(Window* W, char c); whatever wmove's and waddch's you've done don't actually show up on the screen until you call the wrefresh function. int wrefresh(WINDOW *win);

Note the calls to usleep(tus) (defined in unistd.h), which is like the sleep(s) function we've used before, except with finer resolution. The argument is the tiume to sleep given in micro-seconds.

Part 1b: debugging pro-tip

Suppose in the Part1 code we wanted to print out an debugging message "Printed a 2!" just after writing the "2". We'd just add the line

cout << "Printed a 2!" << endl;

... immediately after the waddch(W,'2');. Compile and run that. What you will find is that this message pops up in the middle of our ncurses screen and messes things up. The "3" is no longer printed immediately to the right of the "2". It's a real problem when your debuggin messages cause new bugs! So how can we write out debugging messages in an ncurses program?

Use the standard error output stream (cerr) instead of standard out (cout) for the debugging message. This doesn't solve the problem immediately, but it helps. To see how ...
Create a fifo in the same directory as your lab work using the command
```
mkfifo err
```
Do an ls and you should see "err" show up in the directory listing. The fifo "err" is like a wormhole in the Unix filesystem. Open a separate terminal window, cd to the lab directory containing "err", and give the command
```
tail -f err
```
Go back to your original terminal and type something like this:
```
echo "whoooooooo" > err
```
You should see the whoooooo pop up in the other terminal. Spooky, eh?
Now run lab11v1 in the original terminal like this:
```
./a.out 2> err
```
... and watch as your ncurses window is pure and unblighted by debugging messages, but the message does pop up at the appropriate time in the second window. What's going on here is this: the 2> err is telling the shell to redirect standard error (cerr in our program) output to the fifo err, from whence it travels by wormhole to the less -f running in the other window.

Note: This debugging technique is really useful!

Part 2: coming and going (lab11v2.cpp)

Write a program that reads in input from the user like this

4
a (10,15)
x (14,29)
k (5,5)
x (18,37)

The flashing cursor can be pretty annoying. If you want to get rid of it (make the cursor invisible) use the curs_set function like this:

curs_set(0);

Note that this call must come after the call to initscr().

before going into ncurses mode and then, once in ncurses mode, draws these to the screen one at a time with pauses in between, and then erases them from the screen one at a time (erase means writing a ' ' in the appropriate position. Note: I fully expect that this will be done with a good design that makes use of structs --- probably more than one --- and functions, and that the solution would work for a wide variety of inputs, not only for the input shown here.

Part 3: You got to move it, move it (lab11v3.cpp)

Now we will animate things, i.e. each character will move on the screen. To start things off, let's have a single character, and let's have it start at row 15, column 30. The character will have a direction associated with it — north, south, east or west. Note: design this program keeping in mind that ultimately there will be many characters moving simulatneously. Your program will consist of a loop:

do {
  // draw character
  // sleep for 100,000 usecs
  // use wgetch to see if the user has pressed 'q'
  // move character one step in its current direction
}while('q' has not been pressed);

Of course at this point your character moves off the screen and things are not that interesting. So, to add a little drama let's say that at each step, right before actually moving, the character has a 1-in-10 chance of randomly turning (i.e. changing its current direction) before moving, with a 50% chance of turning 90 left and a 50% chance of turning 90 degrees to the right.

Part 4: Momma should I build a wall? (lab11v4.cpp)

Though your single animated character makes random turns, eventually it's going to make its way off the screen. It'd be much cooler if the screen acted like a walled-in space that the character cannot escape. So let's do that. The only real difficulty here is that you need to know the height and width of the terminal window (in characters). Ncurses has a functin for that:

int getmaxyx(Window *W, int &row, int & col);

So the call getmaxyx(W,height,width) sets height to the number of rows on the screen, and width to the number of columns. Note, then, that the valid positions you can wmove(W,row,col) to are when 0 ≤ row < height and 0 ≤ col < width.

Now that you can get the height and width of the terminal window, when you come to a "move" step you must check to see whether the move in question would take you off the screen. If it does, then instead of moving change direction and leave it at that. You can either randomly change direction, or "bounce" by simply reversing direction (north goes to south, east to west, etc). If you do this, your character will be walled in and will never leave the terminal window.

Overview

Part 1: ncurses basics (lab11v1.cpp)

Part 1b: debugging pro-tip

Part 2: coming and going (lab11v2.cpp)

Part 3: You got to move it, move it (lab11v3.cpp)

Part 4: Momma should I build a wall? (lab11v4.cpp)

Part 5: Yes it's that easy (lab11v5.cpp)

Part 6: Imperio (lab11v6.cpp)

Part 7: And in the end ... (lab11v7.cpp)