Thread: Exercise 1-18 from "The C programming language 2nd edition"

I propose that your algorithm is too complicated. Consider that you only need to remove trailing blanks, and this is arguably true even if the line is empty save for the blanks. Therefore, a simpler algorithm would be:

Code:

search from the end of the line to find the last non-blank character
if no such character is found
    set the string to be an empty string
else
    set the character after the last non-blank character to be a null character

By searching from the end (i.e., first find the null character that terminates the string... or if you use fgets instead of reading character by character this could be the newline character that you then replace with a null character), the first non-blank character encountered is therefore the last non-blank character in the string. What you have to be careful about is to avoid accessing the array out of bounds from the front in the event all the characters are blank, but as long as you think about loop indices carefully that should be fine.

He already has a reasonable way of determining where any terminal run of spaces starts.

What is complicating the function is that it's trying to do too much. There's no reason for it to signal if the line is blank or not. The main function can determine if the line is empty or not by checking the first char.

There's also no reason to try to signal the EOF condition at the same time that you're returning valid data. The getchar will return EOF next time it's executed, too, so just leave the EOF detection until then.

Code:

// main
    while (trim(line, size))
        if (line[0] != '\0')
            printf("%s\n", line);

// trim
    for (i = 0; i < size-1 && (c = getchar()) != EOF && c != '\n'; ++i) {
        s[i] = c;
        if (c == ' ' || c == '\t') ++blanks_count;
        else blanks_count = 0;
    }
    if (c == EOF && i == 0)
        return 0;
    s[i - blanks_count] = '\0';
    return 1;

Originally Posted by john.c

He already has a reasonable way of determining where any terminal run of spaces starts.

What is complicating the function is that it's trying to do too much. There's no reason for it to signal if the line is blank or not. The main function can determine if the line is empty or not by checking the first char.

There's also no reason to try to signal the EOF condition at the same time that you're returning valid data. The getchar will return EOF next time it's executed, too, so just leave the EOF detection until then.

Code:

// main
    while (trim(line, size))
        if (line[0] != '\0')
            printf("%s\n", line);

// trim
    for (i = 0; i < size-1 && (c = getchar()) != EOF && c != '\n'; ++i) {
        s[i] = c;
        if (c == ' ' || c == '\t') ++blanks_count;
        else blanks_count = 0;
    }
    if (c == EOF && i == 0)
        return 0;
    s[i - blanks_count] = '\0';
    return 1;

Thanks.
This version is less complicated and elegant. I will try not to create a super trim program at this moment anyway.
How I understand if 'i' is equal to zero we should exit 'while' loop. So if we type on the terminal enter this will produce a sting containing only '\n' and the value of 'i' will be zero (and as consequence 'return 0' from trim routine). But program does not exit the 'while' loop.
Edit. I forgot that there is an '&&' with 'EOF'. It's clear.