id	summary	reporter	owner	description	type	status	milestone	component	version	severity	resolution	keywords	cc
8407	Unclear docs for is_base_of	driscoll@…	John Maddock	"It isn't clear (at least to me) from [http://www.boost.org/doc/libs/1_53_0/libs/type_traits/doc/html/boost_typetraits/reference/is_base_of.html the documentation for is_base_of] whether it evaluates to `true_type` when `Base` is an //direct// base class of `Derived` or //any// inheritance ancestor. It would be nice if this were clarified.

(I understand the first para of chapter 10 of the C++03 standard to define the term ""base class"" as including both direct and indirect base classes, but IMO this is surprising; myself and two nearby people I informally polled equate ""base class"" with ""direct base class"". At the suggestion of a fourth, we looked it up in the standard for the purposes of this report.)

I had to try the following program to figure out what the behavior is (which seems to be ""`true` if `Base` is any ancestor of `Derived`""):
{{{
#include <boost/type_traits.hpp>
#include <iostream>

struct Base {};
struct Intermediate : Base {};
struct Derived : Intermediate {};

template <typename Base, typename Derived>
const char * answer(Base a, Derived b) {
    return boost::is_base_of<Base, Derived>::value ? ""yes"" : ""no"";
}

int main() {
    Base b;
    Intermediate i;
    Derived d;   
    std::cout << ""Checks:\n""
              << ""  Base is a base class of base [expect yes]:      "" << answer(b, b) << ""\n""
              << ""  Inter. is a base class of base [expect yes]:    "" << answer(b, i) << ""\n""
              << ""  Derived is a base class of inter. [expect yes]: "" << answer(i, d) << ""\n""
              << ""  Base is a base class of derived [expect no]:    "" << answer(d, b) << ""\n""
              << ""Actual answer:\n""
              << ""  Derived is a base class of base: "" << answer(b, d) << ""\n"";
}
}}}"	Bugs	closed	To Be Determined	type_traits	Boost 1.53.0	Problem	fixed