GitLab

It seems like the intended solution to wc is

while (!feof(f)) {
    fread(result, 1, 1, f);
    if (result[0] == '\n') {
        num_lines++;
    }
}

But I think in C it's more idiomatic to check the return value of file reading functions like fread than to use feof. The reason is feof returns true only if you read past end of file using another function. Thus the last fread will read nothing and not update result[0], so the last result[0] will be checked twice. This leads to a bug: if the last character in a file is \n, the program will actually count 1 more than the number of newlines. (Thus it will count 2 lines for data/one and 3 lines for data/two.) Also see here: https://stackoverflow.com/a/12337620.

Here's a version which uses the return value of fread and doesn't double count.

while (fread(result, 1, 1, f) == 1) {
    if (result[0] == '\n') {
        num_lines++;
    }
}

The wc they write should match what the real wc does which is what the current choice does, I believe.

But wc -l data/one returns 1 and wc -l data/two returns 2.

i'm not sure what to do about this, because i really don't want to introduce that pattern at the beginning; @mewu @csander thoughts?

Could they just do this whole assignment with getc() instead? Then they can just compare its return value with EOF.

Although I guess that defeats the purpose of having to allocate a buffer to store the result.

the other thing is my implementation passed; i think you just need to increment in a different place to ignore the last newline

Or they could just put the feof() call after the fread().

yeah that works too. let's just leave it the way it is...hopefully it won't be too confusing, but it'll make for a nice "why is this output wrong" "how do i check" moment.

closed

I think this is going to be quite confusing for them to debug. I would infer from "returns non-zero if we've run out of characters to read in the file and 0 otherwise" that it works like hasNext(). If I noticed the line numbers were off, I would think it was a bug on my part, not a misunderstanding of how the functions work. I would do one of the following:

• Remove the description of feof() and just say that the return value of fread() tells you how many characters were read (which may be less than number_to_read if the end of the file is reached).
• Change the description of feof() to say that it returns non-zero if the last call to fread() tried to read past the end of the file.

reopened

i edited the description to match your second suggestion. let me know if you're happy with it now. :)

Looks good to me

closed

awesome

reopened

closed