为什么该程序产生错误的输出

Question

This program is supposed to remove all comments from a C source code (in this case comments are considered double slashes '//' and a newline character '\\n' and anything in between them, and also anything between '/* ' and '*/'.

The program:

#include <stdio.h>

/* This is a multi line comment
testing */

int main() {
int c;

while ((c = getchar()) != EOF)
{
    if (c == '/') //Possible comment
    {
        c = getchar();
        if (c == '/') // Single line comment
            while (c = getchar()) //While there is a character and is not EOF
                if (c == '\n') //If a space character is found, end of comment reached, end loop
                    break; 


        else if (c == '*') //Multi line comment 
        {
            while (c = getchar()) //While there is a character and it is not EOF
            {   
                if (c == '*' && getchar() == '/') //If c equals '*' and the next character equals '/', end of comment reached, end loop
                    break;
            }                   
        }

        else putchar('/'); putchar(c); //If not comment, print '/' and the character next to it
    }

    else putchar(c); //if not comment, print character
}

}

After I use this source code as its own input, this is the output I get:

#include <stdio.h>

* This is a multi line comment
testing *

int main() {
int c;

while ((c = getchar()) != EOF)
{
    if (c == '') ////////////////
    {
        c = getchar();
        if (c == '') ////////////////////
            while (c = getchar()) /////////////////////////////////////////
                if (c == '\n') ///////////////////////////////////////////////////////////////
                    break; 


        else if (c == '*') ///////////////////
        {
            while (c = getchar()) ////////////////////////////////////////////
            {   

No more beyond this point. I'm compiling it using g++ on the ubuntu terminal.

As you can see, multi lines comments had only their '/' characters removed, while single line ones, had all their characters replaced by '/'. Apart from that, any '/' characters that were NOT the beginning of a new comment were also removed, as in the line if (c == ''), which was supposed to be if (c == '/').

Does anybody know why? thanks.

Answer 1

C does not take notice of the way you indent your code. It only cares about its own grammar.

Look carefully at your else s and think about which if they attach to (hint: the closest open one).

There are other bugs, as well. EOF is not 0, so only the first while is correct. And what happens if the comment looks like this: /* something **/ ?

Answer 2

You have some (apparent) logic errors...

1.

while (c = getchar()) //While there is a character and is not EOF

You're assuming that EOF == 0 . Why not be explicit and change the preceding line to:

while((c = getchar()) != EOF)

2.

else putchar('/'); putchar(c);

Are both of the putchar s supposed to be part of the else clause? If so, you need braces {} around the two putchar statements. Also, give each putchar its own line; it not only looks nicer but it's more readable.

Conclusion

Other than what I've mentioned, your logic looks sound.

Answer 3

As already mentioned, the if/else matching is incorrect. One aditional missing functionality is that you must make it more stateful to keep track of whether you are inside a string or not, eg

printf("This is not // a comment\n");