How to force user to input a value in the type limit

Hi guys,

I read this brilliant post by hammer about, how to get numbers safly using strtol() function.

My question is, how can I make sure the user DO NOT exceed the type limit??

Code:

if (fgets(buf, sizeof(buf), stdin) != NULL)
  {
    i = strtol(buf, &p, 10);

    if (buf[0] != '\n' && (*p == '\n' || *p == '\0'))
      printf ("Valid number of %ld entered\n", i); 
    else  printf ("Invalid number entered\n");
  }

If the user enter 91234567890 that will not be the desirable result becuase the value is bigger long maximum value.

Is there anything I can do to avoid this problem?

Many thanks....

http://www.hmug.org/man/3/strtol.php

Code:

The strtol(), strtoll(), strtoimax() and strtoq() functions return the
     result of the conversion, unless the value would underflow or overflow.
     If no conversion could be performed, 0 is returned and the global vari-
     able errno is set to EINVAL.  If an overflow or underflow occurs, errno
     is set to ERANGE and the function return value is clamped according to
     the following table.

           Function       overflow    underflow
           strtol()       LONG_MIN    LONG_MAX
           strtoll()      LLONG_MIN   LLONG_MAX
           strtoimax()    INTMAX_MIN  INTMAX_MAX
           strtoq()       LLONG_MIN   LLONG_MAX

Basically, check if errno == ERANGE after calling strtol()

Excellent INPUT!!!! Thank you very much.

When using errno in a loop, do I have to asssign it to 0 every round?

Code:

for (;;)
{
    . . .
    errno=0;
    i = strtol(buf, &p, 10);
    if (errno==ERANGE)
        puts("Out of range!");
    else
        break;
}

I though errno will be assigned a new value every time I call strtol() but it doesn't, when error occur , errno keeps the error even when a valid input entered after, is that how is suppose to be?

THANK YOU!!!!

Originally Posted by salvadoravi

is that how is suppose to be?

THANK YOU!!!!

Yes. errno only gets set when an error happens, not when an error fails to happen. (This way things like assigning a value to i, which works, don't keep you from noticing the error from the call to strtol.)

The "variable" errno is not guaranteed to be settable - it is commonly implemented as a macro that calls a funciton that returns the value - this is so that errno can be threadsafe.

It is probably better to do something like this:

Code:

if (result == LONG_MAX  || result == LONG_MIN)
{ 
    if (errno == ERANGE)
        ... value out of range 
}

--
Mats

Originally Posted by matsp

The "variable" errno is not guaranteed to be settable - it is commonly implemented as a macro that calls a funciton that returns the value - this is so that errno can be threadsafe.

It is probably better to do something like this:

Code:

if (result == LONG_MAX  || result == LONG_MIN)
{ 
    if (errno == ERANGE)
        ... value out of range 
}

--
Mats

I see... One question, though: what if the user entered LONG_MAX or LONG_MIN as a value, after errno has been set to ERANGE by a previous iteration? Should we "sacrifice" those two values in order to ensure we're not misusing "errno"?

Sorry, I'm wrong - errno is guaranteed "to be a modifiable lvalue", so yes, you can set it to zero if you like. [It may still be a function call, for example

Code:

extern int *ferrno();
#define errno *ferrno()

would do that. ]

--
Mats

Can you set errno to zero prior to the strtol() call?

Thank you again guys, great words!!!

Originally Posted by Todd Burch

Can you set errno to zero prior to the strtol() call?

Yes, that's the suggested solution.

--
Mats

Originally Posted by matsp

The "variable" errno is not guaranteed to be settable - it is commonly implemented as a macro that calls a funciton that returns the value - this is so that errno can be threadsafe.

--
Mats

I don't know if it was different in C89, but C99 says this:

Originally Posted by C99 Standard, 7.5 Paragraph 2

The macros are
EDOM
EILSEQ
ERANGE
which expand to integer constant expressions with type int, distinct positive values, and
which are suitable for use in #if preprocessing directives; and
errno
which expands to a modifiable lvalue that has type int, the value of which is set to a
positive error number by several library functions. It is unspecified whether errno is a
macro or an identifier declared with external linkage.

So at least in newer implementations, errno should be settable. In fact, footnote 170 states that a program that wishes to check errno should set it to zero before the library call, and check it afterwards.

Originally Posted by tabstop

So at least in newer implementations, errno should be settable. In fact, footnote 170 states that a program that wishes to check errno should set it to zero before the library call, and check it afterwards.

I don't think it's changed any time real recently, I was just confusing it with the fact that you should not declare errno as a variable, as that CAN certainly make a mess of things.

--
Mats