A little problem about cin

[Rene]

I am writing a program about clock these days.
but there are a little problem about input, then i have writen a test program.And i have found a strange thing.
This is my code:
[cpp]
#include <iostream.h>

void main()
{
int a;
while (cin >> a)
cout << a << endl;
}
[/cpp]

• input:
  15
  00
  01
  02
  05
  07
  08
  09
  10

• output:
  15
  0
  1
  2
  5
  7
  0
  8
  0
  9
  10

The question is:why input 01,02,05 and 07(in fact all input small than 08) a = 1,2,5,7.... But when input is bigger than 07,a = 0,8 and 0,9.
I have input only one number,why a have two number?
Please help me,Thank you.

[Julien Cornebise]

Hi
Wich compiler do you use ? Gcc 3.2 ? 2.96 ? TC ? VC++ ? on wich os ? Linux ? Windows ?
Have you tried to include <iostream> and not <iostream.h> (wich is older and now provided for backward compatibility, though I don't think it changes much) ?

[Rene]

I use TC and VC++. They have the same problem.
My OS is Windows.
can i include <iostream>?
i always use iostream.h.
Thank you.

[Viktoras Jucikas]

can i include <iostream>?

You should include iostream, not iostream.h, as .h are provided only for backwards compatibility (for compilers which do not support namespaces, as everything in iostream resides in std namespace). For old c inludes like math.h you should use cmath.

The question is:why input 01,02,05 and 07(in fact all input small than 0 a = 1,2,5,7.... But when input is bigger than 07,a = 0,8 and 0,9.
I have input only one number,why a have two number?

The problem is that any number with leading '0' is treated as an octal number. Thus 08 and 09 are not valid numbers... Stream routines read until they find 8 or 9 after leading 0, then stop and output number they have read (this is 0 in your case). After that they read further and find perfectly valid decimal numbers 8 or 9.

So in this case you're probably better off with reading char[3]...

[Julien Cornebise]

I don't know much in C++ (not yet, I'm learning too), but if reading "09" from input is really a pain in the neck, you still can use scanf() function, wich is originally used in C and is replaced in C++ with stream management, but can still be used.
I personnally chose this for ACM problems, because I've been using C for 2 years and am only coming to C++, so I know printf and scanf kinda well, and much better than I know C++ streams.

[Rene]

yes,i can use scanf(). but i don't know much about the function in C. In UVA, the most input is terminated by End Of File. when i use stream, cin will throw an exception when it read EOF, i use
[cpp] while (cin >> a)[/cpp] to find the End Of File.
If i use scanf(), how should i write
[c]while (scanf ("%d",a) != EOF)[/c]
is it right?

[Julien Cornebise]

Well, the thing that works the most is to use the fact that scanf returns the number or arguments correctly read :
[c]
int a, b;
scanf("%d %d", &a, &b);
[/c]
will return 2 if a AND b has been read, separated by a space.
So the easer thing to do is to
[c]while( scanf("%d", &a) == 1) [/c]
And that should be OK (that's what I've done so far, from the beginning).

[Krzysztof Duleba]

Let T be a type. cin>>T, if defined, reads the biggest continuous string of characters and makes it to fit T type. So when T is a char, you read only one character. When T is a number type, you read the whole number.
And a string is continuous if it doesn't contain any spaces or end of file/line characters. Such characters are not removed from the stream, except for the case with T being a char(or array of chars I suppose - I don't know as I use string template class instead )