[math] regression test failure on GCC 4.5 in c++0x mode

[Richard Webb <richard.webb@…>]

When running the math regression tests in Mingw 4.5.0 in C++0x mode, i get some failures due to ambiguities between functions in boost::math and std::. e.g.

​http://tinyurl.com/2blmq3l ​http://tinyurl.com/26yevqq

qualifying the calls with boost::math allows the tests to compile.

[Paul A. Bristow]

Your fix is fine, and may be best for some applications, but we feel it reads verbosely.

So we have decided to recommend a different way in our examples.

We recommend *NOT* "using namespace anything;" std, boost::math ...

but to qualify each distribution used after the

#include <boost/math/my_distribution>

like

"using boost::math::my_distribution;"

(and "using std::cout; ..." if you wish - worthwhile clutter reduction if there are lots of calls to cout, as is common in cpp source files, but calls are less common in .hpp include files.)

This can be done globally or locally in source .cpp files, but should be only locally in .hpp include files.

All our examples will now follow this recommendation.

[John Maddock]

Closing down, this should now be fixed.