Please ask this on the libstdc++ list instead. On 1 August 2013 23:37, niXman <i.nixman@xxxxxxxxx> wrote: > Hi, > > Some time ago, I thought about adding CryptoAPI[1] support for > 'std::random_device'. Now, I have some free time, and I want to do it. > But, after the analysis of the 'libstdc++-v3/include/bits/random.h' > and 'libstdc++-v3/src/c++11/random.cc' files, I had a few questions. > Firstly, when building gcc-trunk using MinGW, the > _GLIBCXX_USE_RANDOM_TR1 macro is not defined (I couldn't understand > was it right or not?). Thus, the following methods are used: > 'void random_device::_M_init_pretr1(const std::string& token)' > 'random_device::result_type random_device::_M_getval_pretr1()' > > The 'random_device::_M_getval_pretr1()' method, currently, as a source > of pseudo-random numbers is using the '_M_mt' member of the > 'std::mt19937' type. And here I do not understand why this is > happening, is it right, and why it's done that way? > > Or maybe, I configure GCC incorrectly, and the _GLIBCXX_USE_RANDOM_TR1 > macro must be defined? > > Thanks! > > > [1] http://msdn.microsoft.com/en-us/library/windows/desktop/aa380255(v=vs.85).aspx > > > -- > Regards, > niXman > ___________________________________________________ > Dual-target(32 & 64-bit) MinGW compilers for 32 and 64-bit Windows: > http://sourceforge.net/projects/mingwbuilds/ > ___________________________________________________ > Another online IDE: http://liveworkspace.org/
