2009/11/11 Emanuele Rocci
HI Roman! Thanks for your nice response! I have just discovered another way. Please give me your comments!
constant_type< const char[7] >::type _name(constant("value: "));
playing the same role of
boost::function
_name = boost::lambda::constant("value: "); I find both 2 nice ways of expressing constant string in a lamda expression.
Using the boost::function approach, some sample code can be written like the following using round brackets next to _name variable.
std::for_each( myList.begin(),myList.end(),( std::cout << _name() << _1 , _1 = 2 ));
This does not do what you think it does. It'll evaluate and output _name() only once. Here's how you can fix it: std::for_each( myList.begin(),myList.end(),( std::cout << bind(_name) << _1 , _1 = 2 ));
I find the approach using boost::function very nice since it provides more the "genericity" feeling; you don't have to declare the size of the string but it is enough const char*().
With the constant_type way it is possible to use _name as a common const variable without using round brackets
std::for_each( myList.begin(),myList.end(),( std::cout << _name << _1 , _1 = 2 ));
This works as expected -- _name is being output several times.
This works fine provided that the size is correct; it is probably not always nice, but compiler will complain if the size is not correct i.e. if you use const char[6] or if you use const char[8].
Now I am wondering which is the fastest approach. I guess that probably they are the same.
Using constant_type should be faster, because it does not use virtual functions. Also, to my surprise, it's a documented way of creating lazy constants (see http://www.boost.org/doc/libs/1_40_0/doc/html/lambda/le_in_details.html#lamb... ). Roman Perepelitsa.