Why I have to cin.ignore() ?

Hello there

Code:

---

#include <iostream>


using namespace std;


int multiply(int x, int y);


int main()
{
    int x,y;
    cout << "Insert the value of x: ";
    cin >> x;
    //cin.ignore();
    cout << "Insert value y: ";
    cin >> y;
    //cin.ignore();
    cout << "X multiple Y equals in " << multiply(x,y) << endl;


    return 0;
}


int multiply(int x, int y) {
    return (x*y);
}

---

Comment or not, cin.ignore(); does nothing in this program output. I read somewhere that we use this function to throw <Enter> away, but I don't understand what does it means... I suppose Enter represents a Carriage Return CR character or something like that, but again... I can't figure out the meaning of use the cin.ignore() since the program output is identical the same using it or not. :confused:

Run your program without the ignore() functions and type "4 7" at the first prompt.

Quote:

---

Originally Posted by tabstop

Run your program without the ignore() functions and type "4 7" at the first prompt.

---

Code:

---

[drpaneas@fedora Debug]$ ./cprogram1 Insert the value of x: 4 7
Insert value y: X multiple Y equals in 28

---

Now do you see the use? (If not, do the same experiment on the "with ignore" version.)

Quote:

---

Originally Posted by tabstop

Now do you see the use? (If not, do the same experiment on the "with ignore" version.)

---

with ignore version:

Code:

---

[drpaneas@fedora Debug]$ ./cprogram1
Insert the value of x: 4 7
Insert value y: X multiple Y equals in 28

---

it's the same :/

EDIT: I suppose cin.ignore doesn't work on GNU GCC compiler ?

Ah, right: ignore by default only ignores one character (which in this case is the space). Something like ignore(80) may do the trick.

Nope, ignore(80) doesn't make my day :p
the program hangs up and waiting until Ctrl+C is pressed manually by the user.

From Wikipedia: "many programming languages treat carriage return and line feed as whitespace."

So cin.ignore ignores the whitespace between 4 7 ?
I suppose then the result would be 47 (one number) instead of 4*7=28. Hmmm ?

You don't need Ctrl-C: you just need to type 80 characters. Have you, at any point, actually looked up what ignore does?

Quote:

---

Originally Posted by tabstop

You don't need Ctrl-C: you just need to type 80 characters. Have you, at any point, actually looked up what ignore does?

---

At last,I think I figured it out :D

Extracts characters from the input sequence and discards them.
The extraction ends when n characters have been extracted and discarded or when the character delim is found, whichever comes first. In the latter case, the delim character itself is also extracted.

So cin.ignore(2); works for me :)

Code:

---

[drpaneas@fedora Debug]$ ./cprogram1 Insert the value of x: 4 7
Insert value y: 3
X multiple Y equals in 12

---

Thanks tabstop \m/

Well I made a post at my blog clarifying things pretty much :D
Why do I have to use cin.ignore(); ? | Dr.Paneas Blog

Quote:

---

Originally Posted by Dr.Paneas

Well I made a post at my blog clarifying things pretty much :D
Why do I have to use cin.ignore(); ? | Dr.Paneas Blog

---

Nice post.
Also, as the post is somewhat long, you could put some other types of examples, like how to ignore a variable number of characters upto a certain one or a maximum number of them as available....etc.

However, you should know that it isn't possible as far as I am aware to protect against all stupidity using the cin >>. Plus such solutions are often complex.
The safest (and best) solution is to always read in the entire line of user input, and then parse it properly to get what you need.
If the user enters a non-integer, the stream will go into an "error state," which is extremely difficult to solve. Refer to this post for more info: "Flushing" the input stream

Quote:

---

If the user enters a non-integer, the stream will go into an "error state," which is extremely difficult to solve.

---

cin.clear();

OK, true, it does clear the error state of the stream, but not the application. The application must, after clearing the error state of the stream, get rid of unwanted left over input in the input stream, or the stream will go into an error state again.
And clearing the left over input in the stream turns out not to be so easy.

Quote:

---

Originally Posted by Elysia

OK, true, it does clear the error state of the stream, but not the application. The application must, after clearing the error state of the stream, get rid of unwanted left over input in the input stream, or the stream will go into an error state again.
And clearing the left over input in the stream turns out not to be so easy.

---

Well, one generally tries to cover all possible types of input during an iteration.
So, after getting an error state and cleaning that, the data may be put into the right place by type, where it may be ignored.

Again, see "Flushing" the input stream

Quote:

---

Originally Posted by Elysia

Again, see "Flushing" the input stream

---

I think you're making mountans out of molehills right now.

Prelude's articles tend to be complete. I don't think that means input is an insurmountable problem. A lot of detail goes into working with wide streams, and fixing other weirdness such as what happens if there is no input waiting in the stream when you ignore. These details are not relevant all of the time. I could plan my input so that such special cases never occur.

Additionally, I would handle EOF all of the time, sure, but that's well defined in the standard.

I merely pointed out that it's not always so "simple," to just put some cin.ignore() statements in well defined points in the code, as the article seems to indicate.

Quote:

---

Originally Posted by Elysia

I merely pointed out that it's not always so "simple," to just put some cin.ignore() statements in well defined points in the code, as the article seems to indicate.

---

That was initially my point anyway !

And you turned my simple correction into a debate. I didn't have any problem with what you said other than what I quoted. Fixing the stream state is easy, after which you'd call some version of ignore_line() from the article. After reading the article, of course.