tokenizing a string made of "only" seperator characters

[aryan1]

Hi All,

Following function utilizing boost's string tokenizer class crashes as might be expected:

bool CheckCommand(const string& cmd, const string& line)
{
	if (line.length() == 0)
		return false;

	typedef tokenizer< char_separator<char> > StrTokenizer;
	char_separator<char> sep(":");
	StrTokenizer tokens(line, sep);
	StrTokenizer::iterator it = tokens.begin();

	if (*it == cmd) { <<<<--------------- Crash takes place at this line
	    ++it;
	    if (it != tokens.end()) {
	        return true;
        }
    }

	return false;
}

when line is "::::::::::::::::::::", i.e. made of merely separator characters (":").

How can I check

StrTokenizer::iterator it

actually points to something meaningful to prevent this crash ?

Thanks.

[superbonzo]

if ( it != tokens.end() && *it == cmd )

bye!