different result from file reading and standard input with getline

Question

bool validateCurrencyLine(string line){
    cout << "TESTING LINE : " << line << endl;
    string pattern = "[ ]*([A-Z]{3}) ([0-9]+)([ ]*|,[0-9]+[ ]*)";
    boost::regex expr{pattern};
    return boost::regex_match(line,expr);
}

int main()
{
    string line;
    while(getline(cin,line)){
        cout << validateCurrencyLine(line) << endl;
    }
    return 0;
}

The content of test file is as follows:

IDK 3453443

Now when I start a program using ./a.out < test the result is

TESTING LINE : IDK 3453443
0
TESTING LINE : 
0

My assumption is that the second line is printed because the testfile first line is actually

IDK 3453443 + enter

Am I correct?)

But the real problem is that when I start it like this: ./a.out and input "IDK 3453443" and press enter. The result of this is:

TESTING LINE : IDK 3453443
1

Any thoughts why these two results differ?

Answer 1

Indeed the line ends are the culprit.

Look at the file in a hex editor, you will find 0d 0a line ends (Windows or CRLF), where the code expects UNIX line-ends (just LF).

See it live:

Live On Coliru

You can work around it by 'eating' all whitespace at the end:

Live On Coliru

std::string pattern = "[ ]*([A-Z]{3}) ([0-9]+)(,[0-9]+)?\\s*";

Now both get accepted.