it is a fix/workaround for boost.thread and boost.exception. all it changes is to use boost::enable_if instead of enable_if Isn't better to understand why ? It's not normal that enable_if inside a nested namespace needs to have boost:: specified and while we are at it why boost::enable_if and not ::boost::enable_if ? feel free to look into it:
* check out trunk * go to libs/lockfree/tests * change the files: move the inclusion of boost/test/unit_test.hpp to the front of the file (or apply attached patch) * compile the testsuite * get a deeper understanding of the situation
i'm not planning to dig deeper into these libraries, myself
As I said, there is already an issue reported on this ML
http://boost.2283326.n4.nabble.com/Boost-Test-g-compilation-error-due-to-amb...
great to see that it is a known issue. not so great to see that the bug is in the tracker for several months without activity ...