Thread: Compiler independent wtoi function

I'm wondering whether if there's a c++ standard counterpart of the MSVC++ _wtoi function which would be compiler independent.

_wtoi works fine for me at the moment but I'm going to be compiling my program on a linux machine with possibly gcc later on and I don't want to hit a snag on this score.

Or should I pursue this workaround that still isn't working

Code:

	unsigned integer i;
	wistringstream wss;
	wstring ws;

	ws = some_wide_character_that_I_get_from_elsewhere;
	wss.str(ws);
	wss >> i;

If _wtoi works regardless of the compiler/system, I'll go with it. But if not, advice would be appreciated.

what about atoi()?

Wouldn't that be susceptible to undefined behavior if you through a two byte character at it? I'm not worried about whether it works on my XP. I'm worried about whether it'll break apart when it goes to gcc or linux.

I'm getting input from a unicode text file. I don't know if there's a way out of that.

You could also have a look at boost::lexical_cast

Thanks for the replies.

What problems are you having? Other than the use of "integer" instead of int, that code works for me in a simple test app.

In my original version, I wasn't actually putting a wchar_t variable into the wstring. I was iterating through a wstring with a wstring::iterator and when it came across a number, it would directly dump it into a temporary wstring. That somehow didn't work so I just dumped the iterator into a wchar_t and then dumped that into a wstring which again went into a wistringstream and now it's working.
I know I should understand why this is working and the other isn't but I'll put that off to another day.

well, here's a way to turn a single wchar_t into a character string with the same byte sequence:
Code:

Code:

#include <iostream>
#include <cstdlib>

int main()
{
	union switchitup
	{
		wchar_t bigletter;
		char smalletters[sizeof(wchar_t)/sizeof(char)];
	} abcd;

	abcd.bigletter=0x0035;	
	std::cout<<atoi(abcd.smalletters)<<std::endl;

	return 0;
}

Thanks. Since I only need the numerical characters, I think this might be the better solution.