ways to validate float point datatypes?

[Sidro]

Hi,
What are some of the ways to validate floating point numbers from a user input? I looked into the functions in c++
but it includes the scientific notation "e", which I dont like.Thanks i advance.

 

[ArkM]

If you don't want to write your own (and very fast double constant parser (very interesting job, try this approach:

bool Validate(const char* s, double& x)
{
  if (!s || !*s)
    return false;
  istrstream val(s,strlen(s));
  val >> x; // let C++ library works...
  return val.good();
}

inline bool Validate(const string& s, double& x)
{
  return Validate(s.c_str(),x);
}

bool Validate(const char* s, double* px = 0)
{
  double x;
  return Validate(s,px?px:&x);
}

inline bool Validate(const string& s, double* px = 0)
{
  return Validate(s.c_str(),px);
}

Alas, old good atof (from <cstdlib>) returns undefined value in overflow case.
C++ streams returns fail() when bad constant occured, but direct stream poll (then clear) is too cumbersome.
Yes, it's not a fastest validator, but an user console is not the fastest input device too...
We need only one Validate() function, but four overloaded functions better than one...