rmaprath added a comment. In http://reviews.llvm.org/D20328#456271, @mclow.lists wrote:
> > I've adopted a initialize-on-first-use policy to workaround this > > particular problem. > > > That's not constexpr. > > You say it in your comment "This prohibits any prospects of calling a runtime > initialization routine", but then you have a runtime initialization routine. `std::mutex()` constructor is still constexpr, what I've done is to defer the initialization to `lock()` and `unlock()` methods. This is pretty much the only way I can keep the constructor constexpr and allow an underlying platform implementation to provide the guts of `std::mutex`. Hope that makes sense? http://reviews.llvm.org/D20328 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
