Thread: Input continuation loop

The problem is that your instructor gave you bad advice: fflush() is only defined for output and update (i.e., both input and output) streams, so fflush(stdin) results in undefined behaviour.

One solution is to replace it with:

Code:

int c;
while ((c = getchar()) != EOF && c != '\n');

You may need to move the definition of c to the start of the while loop. The idea is to read and discard characters until the end of file is reached or the first newline character is reached (the newline would have been from the previous entering of input).

Originally Posted by Sinosizer

We routinely use fflush(stdin) followed by getchar() to prompt the user to type something before our little programs exit. :-(

Then you are relying on undefined behavior. IMHO, any decent C runtime should explode violently when attempting to flush an input buffer, since the very idea makes no sense. Unfortunately, some platforms will do what you want, leading you to believe your code is correct when it actually isn't.

The problem is extra input in the buffer prior to the next newline character. The solution is to consume this input deliberately. It's not too hard:

Code:

void ignoreRestOfLine(FILE *fp)
{
    int ch;

    while((ch = fgetc(fp)) != EOF && ch != '\n')
        ;
}

Originally Posted by Sinosizer

I spoke to my instructor about this, and he suggested the code involving EOF and \n to be used on my Unix code. The fflush function call works somehow in Visual Studio 2008, even though it's not recommended.

My question is, do I augment the EOF and \n portion with

Code:

&& c == 'Y'

Speak to your instructor again. EOF and '\n' are perfectly valid in ALL platforms that support stdio.h - which if you exclude some of the smaller embedded plaforms is just about all computer platforms you will ever encounter (and of course, many programers never have the pleasure of working on a small embedded system that is designed for a specific task).

--
Mats