Lecture 9 – Sep 23rd, 2019

Setup

Log in to clyde.
Create a directory and cd into it.
Copy ~steve/ex/size.c to your directory.

Task

Open size.c in your editor and read it. Try to guess what it will print out. Compile the code and run it.
```
$ clang -Wall -std=c11 -o size size.c
$ ./size
```
Did it print what you expected?
Modify size.c and add additional printf lines to print the sizes of short, int, long, long long, float, double, and size_t. Compile your program and run it.
Modify size.c one more time. This time, print the size of bool. Try to compile the code. Clang gives two errors, the second is nonsensical, but the first is instructive. Try to compile with gcc instead of clang. You’ll get a very similar error.
Look closely at the error messages, they tell you a lot. They start by giving the name of the file, the line of the error, and the column of the error: size.c:line:column:. This is followed by the error message and then the line of code with a marker indicating the place the compiler thinks you’ve made an error.
(You can have Vim show you line numbers by adding set number to your ~/.vimrc or by running :set number while Vim is running. You can have Emacs display line numbers too.)
This particular error is because C calls its Boolean type _Bool. This is terrible. Fortunately, there’s a better way.
Add the line #include <stdbool.h> at the top of size.c. This makes bool an alias (technically a typedef) of _Bool and defines true as 1 and false as 0. Compile and run it.
(If you’d like, you can read /usr/include/clang/6.0/include/stdbool.h to see how that happens, but note we haven’t actually asked about most of what’s in there. Nevertheless, it should be fairly easy to follow what’s going on.)
Run
```
$ clang-format -i size.c
```
to make sure your code is consistently formatted.
Create a new file, fib.c. Write a function long long fib(int n) that computes and returns the nth Fibonacci number. (Feel free to use either recursion or a loop. Make sure that any temporary variables you use have type long long.)
Below it, create a main function that calls fib with the argument of your choice and prints out the result, e.g.,
```
int n = 19;
printf("F_%d = %lld\n", n, fib(n));
```
Format your code with clang-format, compile and run it. Make sure you get the right answer!
Read the man page for atoi. This function takes a string and returns an integer. Modify fib.c to iterate over all of the command line parameters (excluding argv[0] which is the name of the program itself), call atoi(argv[idx]) for each 1 ≤ idx < argc, and print out the result.
Pay attention to which header file you need to include in order to use atoi and make sure to #include it.
Format, compile, and test.
```
$ ./fib 0 19 12
F_0 = 0
F_19 = 4181
F_12 = 144
```
Modify fib.c such that if any of the arguments are negative, it prints out an error message rather than the value of the Fibonacci number. If any of the arguments are negative, make main return 1 to indicate error.
Recall that error messages should go to stderr rather than stdout. We can use fprintf to print to the stderr file.
```
fprintf(stderr, "%s: Error: negative number: %s\n", argv[0], argv[idx]);
```
Example output.
```
$ ./fib 2 -3 6
F_2 = 1
./fib: Error: negative number: -3
F_6 = 8
$ echo $?
1
```