I don't understand "while (getchar() != '\n');"

Hi guys,

I'm one of the many who encountered a problem with the getchar() function picking up the newline '\n' as an input on subsequent getchar() statements.

I googled and found the following solution such that the second getchar() doesn't get a newline as an input and it works.

Code:

---

input = getchar();
while (getchar() != '\n');
input = getchar();

---

However, I don't get how the while (getchar() != '\n'); eliminates any newlines such that the following getchar() doesn't capture the newline as an entry.

I tried searching for this on Google but I didn't find any answers(that I understand) regarding this matter. :/

Can anyone explain, please? Thanks! And please keep in mind I only started programming 4 weeks ago which means I'm like absolutely noobish. :(

getchar fetches a character from the input stream every time it's called. So the loop basically does this:

Call getchar.
Compare getchar to '\n'.
If not equal, break the loop.

Remember that the last character in the input buffer will be the newline, so that's why it works.

Example:
Left in input buffer: Hello World\n
Loop calls getchar.
Left in input buffer: ello World\n
...
Loop calls getchar.
Left in input buffer: \n
Loop calls getchar.
getchar returns '\n', so loop breaks.
Left in input buffer: nothing.

OK, forget about getchar for a moment, just pretend that c is its return value.

while( c != '\n' ) ;

When is this not true? If c=='\n', therefore, getchar read a newline. Under normal circumstances there wouldn't be anything to read after that, so the next getchar waits for input.

Ahh... So getchar() reads from left to right... I was actually sort of thinking it read the string from right to left... XD So that's why. And truly, thanks alot! You guys answered a question that's been ailing me for 2 days in just 10 minutes o.o

Another solution I found involves using scanf:

Code:

---

scanf("%c",&input);
scanf(" %c",&input);

---

How exactly does the space actually eliminate the "\n" entered anyway? o.o I read on Google, and they just say they just "sort of" cancel each other out. Not very clear :/

Format strings can be a mess, but if you can keep it simple they're easy enough to use. In the second statement, the format string has a whitespace character, which communicates to scanf to ignore any and all such characters. Other conversions don't need this kind of treatment. scanf("%d%s", userid, name); would work the same with "%d %s" as the first argument, but with chars, the ignoring of whitespace is suppressed.

this is what a\nmultiline file actually\nlooks like when you access it programatically\n:)

this is what a
multiline file actually
looks like when something like
notepad opens the file
:)

there are many other escape characters, another you might find in a text file is \t for tabs. etc...

IF (TRUE) continue ELSE gtfo.

LINE #1
't' != '\n' - TRUE
'h' != '\n' - TRUE
'i' != '\n' - TRUE
's' != '\n' - TRUE
' ' != '\n' - TRUE
'i' != '\n' - TRUE
's' != '\n' - TRUE
// ...
'\n' != '\n' - FALSE *** GTFO! >:]

LINE #2
't' != '\n' - TRUE
'h' != '\n' - TRUE
// ...
'\n' != '\n' - FALSE *** GTFO! >:]

... until end of file.

Thanks. :D I understand while(getchar() != '\n'); much better now :)

I also ran a test on scanf such that I replace the front part of the input specifier with different whitespace characters.

For example:
scanf("\n%c",&input);
scanf("\t%c",&input);
scanf("\f%c",&input);

and so on...

They seem to do the trick for telling C++ to skip over any and all whitespaces behind the character. I'm assuming that putting a whitespace in front of the input specifier instructs C++ to read all whitespace characters then look for a %c character to assign to the input variable.

Is my assumption right? Or did I misinterpret your explanation wrongly, whiteflags?

And really, many thanks for the help! :)

Don't use scanf in C++. Use cin >>.
To clear the input buffer, do

Code:

---

std::cin.clear(std::numerical_limits<int>::max());

---

(I believe.)

No, no, that's all wrong! :P Never use "int" when you can typedef another more complicated name and use that instead.

Code:

---

std::cin.ignore ( std::numeric_limits<std::streamsize>::max(), '\n' );

---

You also need to include <limits> and <ios> (along with <iostream> if you're clearing std::cin).
http://cboard.cprogramming.com/cplus...tml#post614384

I never use it enough to know what it accepts or the names of arguments of those stuff... so yeah, don't blame my bad memory. I did try...

Quote:

---

Originally Posted by dwks

No, no, that's all wrong! :P Never use "int" when you can typedef another more complicated name and use that instead.

---

The alternative is of course to save 36 characters and use INT_MAX from <climits>.

My point was that std::streamsize isn't guaranteed to be of type int (at least as far as I know).

But yeah, you could just use INT_MAX. That's sufficiently high enough that it would almost always work . . . .

Yeah, I'm aware. Depending on the context, and the urgency of what I'm trying to do, I'd either use the streamsize version or INT_MAX.