// random number generation (out of line) -*- C++ -*-

// Copyright (C) 2006 Free Software Foundation, Inc.
//
// This file is part of the GNU ISO C++ Library.  This library is free
// software; you can redistribute it and/or modify it under the
// terms of the GNU General Public License as published by the
// Free Software Foundation; either version 2, or (at your option)
// any later version.

// This library is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
// GNU General Public License for more details.

// You should have received a copy of the GNU General Public License along
// with this library; see the file COPYING.  If not, write to the Free
// Software Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301,
// USA.

// As a special exception, you may use this file as part of a free software
// library without restriction.  Specifically, if other files instantiate
// templates or use macros or inline functions from this file, or you compile
// this file and link it with other files to produce an executable, this
// file does not by itself cause the resulting executable to be covered by
// the GNU General Public License.  This exception does not however
// invalidate any other reasons why the executable file might be covered by
// the GNU General Public License.

#include <limits>

namespace std
{
_GLIBCXX_BEGIN_NAMESPACE(tr1)

  /*
   * Implementation-space details.
   */
  namespace _Private
  {
    // General case for x = (ax + c) mod m -- use Schrage's algorithm to avoid
    // integer overflow.
    //
    // Because a and c are compile-time integral constants the compiler kindly
    // elides any unreachable paths.
    //
    // Preconditions:  a > 0, m > 0.
    //
    template<typename _Tp, _Tp a, _Tp c, _Tp m, bool _m_is_zero>
      struct Mod
      {
	static _Tp
	calc(_Tp x)
	{
	  if (a == 1)
	    x %= m;
	  else
	    {
	      static const _Tp q = m / a;
	      static const _Tp r = m % a;
	      
	      _Tp t1 = a * (x % q);
	      _Tp t2 = r * (x / q);
	      if (t1 >= t2)
		x = t1 - t2;
	      else
		x = m - t2 + t1;
	    }

	  if (c != 0)
	    {
	      const _Tp d = m - x;
	      if (d > c)
		x += c;
	      else
		x = c - d;
	    }
	  return x;
	}
      };

    // Special case for m==0 -- use unsigned integer overflow as modulo
    // operator.
    template<typename _Tp, _Tp a, _Tp c, _Tp m>
      struct Mod<_Tp, a, c, m, true>
      {
	static _Tp
	calc(_Tp x)
	{ return a * x + c; }
      };

    // Dispatch based on modulus value to prevent divide-by-zero compile-time
    // errors when m == 0.
    template<typename _Tp, _Tp a, _Tp c, _Tp m>
      inline _Tp
      mod(_Tp x)
      { return Mod<_Tp, a, c, m, m == 0>::calc(x); }

    // Like the above, for a==1, c==0, in terms of w.
    template<typename _Tp, _Tp w, bool>
      struct Mod_w
      {
	static _Tp
	calc(_Tp x)
	{ return x % (_Tp(1) << w); }
      };

    template<typename _Tp, _Tp w>
      struct Mod_w<_Tp, w, true>
      {
	static _Tp
	calc(_Tp x)
	{ return x; }
      };

    // Selector to return the maximum value possible that will fit in 
    // @p w bits of @p _Tp.
    template<typename _Tp, _Tp w, bool>
      struct Max_w
      {
	static _Tp
	value()
	{ return (_Tp(1) << w) - 1; }
      };

    template<typename _Tp, _Tp w>
      struct Max_w<_Tp, w, true>
      {
	static _Tp
	value()
	{ return std::numeric_limits<_Tp>::max(); }
      };

  } // namespace _Private


  /**
   * Constructs the LCR engine with integral seed @p x0.
   */
  template<class UIntType, UIntType a, UIntType c, UIntType m>
    linear_congruential<UIntType, a, c, m>::
    linear_congruential(unsigned long x0)
    { this->seed(x0); }

  /**
   * Constructs the LCR engine with seed generated from @p g.
   */
  template<class UIntType, UIntType a, UIntType c, UIntType m>
    template<class Gen>
      linear_congruential<UIntType, a, c, m>::
      linear_congruential(Gen& g)
      { this->seed(g); }

  /**
   * Seeds the LCR with integral value @p x0, adjusted so that the 
   * ring identity is never a member of the convergence set.
   */
  template<class UIntType, UIntType a, UIntType c, UIntType m>
    void
    linear_congruential<UIntType, a, c, m>::
    seed(unsigned long x0)
    {
      if ((_Private::mod<UIntType, 1, 0, m>(c) == 0)
	  && (_Private::mod<UIntType, 1, 0, m>(x0) == 0))
	m_x = _Private::mod<UIntType, 1, 0, m>(1);
      else
	m_x = _Private::mod<UIntType, 1, 0, m>(x0);
    }

  /**
   * Seeds the LCR engine with a value generated by @p g.
   */
  template<class UIntType, UIntType a, UIntType c, UIntType m>
    template<class Gen>
      void
      linear_congruential<UIntType, a, c, m>::
      seed(Gen& g, false_type)
      {
	UIntType x0 = g();
	if ((_Private::mod<UIntType, 1, 0, m>(c) == 0)
	    && (_Private::mod<UIntType, 1, 0, m>(x0) == 0))
	  m_x = _Private::mod<UIntType, 1, 0, m>(1);
	else
	  m_x = _Private::mod<UIntType, 1, 0, m>(x0);
      }

  /**
   * Returns a value that is less than or equal to all values potentially
   * returned by operator(). The return value of this function does not
   * change during the lifetime of the object..
   *
   * The minumum depends on the @p c parameter: if it is zero, the
   * minimum generated must be > 0, otherwise 0 is allowed.
   */
  template<class UIntType, UIntType a, UIntType c, UIntType m>
    typename linear_congruential<UIntType, a, c, m>::result_type
    linear_congruential<UIntType, a, c, m>::
    min() const
    { return (_Private::mod<UIntType, 1, 0, m>(c) == 0) ? 1 : 0; }

  /**
   * Gets the maximum possible value of the generated range.
   *
   * For a linear congruential generator, the maximum is always @p m - 1.
   */
  template<class UIntType, UIntType a, UIntType c, UIntType m>
    typename linear_congruential<UIntType, a, c, m>::result_type
    linear_congruential<UIntType, a, c, m>::
    max() const
    { return (m == 0) ? std::numeric_limits<UIntType>::max() : (m - 1); }

  /**
   * Gets the next generated value in sequence.
   */
  template<class UIntType, UIntType a, UIntType c, UIntType m>
    typename linear_congruential<UIntType, a, c, m>::result_type
    linear_congruential<UIntType, a, c, m>::
    operator()()
    {
      m_x = _Private::mod<UIntType, a, c, m>(m_x);
      return m_x;
    }


  template<class _UInt, int w, int n, int m, int r,
	   _UInt a, int u, int s,
	   _UInt b, int t, _UInt c, int l>
    void
    mersenne_twister<_UInt, w, n, m, r, a, u, s, b, t, c, l>::
    seed(unsigned long value)
    {
      if (value == 0)
	value = 4357;
      
#if 0
      // @todo handle case numeric_limits<_UInt>::digits > 32
      if (std::numeric_limits<_UInt>::digits > 32)
	{
	  std::tr1::linear_congruential<unsigned long, 69069, 0, 2**32> lcg(value);
	  seed(lcg);
	}
      else
	{
	  std::tr1::linear_congruential<unsigned long, 69069, 0, 0> lcg(value);
	  seed(lcg);
	}
#else
      std::tr1::linear_congruential<unsigned long, 69069, 0, 0> lcg(value);
      seed(lcg);
#endif
    }


  template<class _UInt, int w, int n, int m, int r,
	   _UInt a, int u, int s,
	   _UInt b, int t, _UInt c, int l>
    template<class Gen>
      void
      mersenne_twister<_UInt, w, n, m, r, a, u, s, b, t, c, l>::
      seed(Gen& gen, false_type)
      {
	using _Private::Mod_w;
	using std::numeric_limits;

	for (int i = 0; i < state_size; ++i)
	  _M_x[i] = Mod_w<_UInt, w,
	                  w == numeric_limits<_UInt>::digits>::calc(gen());
	_M_p = state_size + 1;
      }

  template<class _UInt, int w, int n, int m, int r,
	   _UInt a, int u, int s,
	   _UInt b, int t, _UInt c, int l>
    typename
    mersenne_twister<_UInt, w, n, m, r, a, u, s, b, t, c, l>::result_type
    mersenne_twister<_UInt, w, n, m, r, a, u, s, b, t, c, l>::
    max() const
    {
      using _Private::Max_w;
      using std::numeric_limits;
      return Max_w<_UInt, w, w == numeric_limits<_UInt>::digits>::value();
    }


  template<class _UInt, int w, int n, int m, int r,
	   _UInt a, int u, int s,
	   _UInt b, int t, _UInt c, int l>
    typename
    mersenne_twister<_UInt, w, n, m, r, a, u, s, b, t, c, l>::result_type
    mersenne_twister<_UInt, w, n, m, r, a, u, s, b, t, c, l>::
    operator()()
    {
      // reload the vector - cost is O(n) amortized over n calls.
      if (_M_p >= state_size)
	{
	  const _UInt upper_mask = (~_UInt()) << r;
	  const _UInt lower_mask = ~upper_mask;

	  for (int k = 0; k < (n - m); ++k)
	    {
	      _UInt y = (_M_x[k] & upper_mask) | (_M_x[k + 1] & lower_mask);
	      _M_x[k] = _M_x[k + m] ^ (y >> 1) ^ ((y & 0x01) ? a : 0);
	    }

	  for (int k = (n - m); k < (n - 1); ++k)
	    {
	      _UInt y = (_M_x[k] & upper_mask) | (_M_x[k + 1] & lower_mask);
	      _M_x[k] = _M_x[k + (m - n)] ^ (y >> 1) ^ ((y & 0x01) ? a : 0);
	    }

	  _M_p = 0;
	}

      // Calculate x(i)
      result_type y = _M_x[_M_p++];
      y ^= (y >> u);
      y ^= (y << s) & b;
      y ^= (y << t) & c;
      y ^= (y >> l);
      
      return y;
    }


  template<typename _IntType, _IntType m, int s, int r>
    void
    subtract_with_carry<_IntType, m, s, r>::
    seed(_IntType __value)
    {
      std::tr1::linear_congruential<unsigned long, 40014, 0, 2147483563>
	lcg(__value);
      
      for (int i = 0; i < long_lag; ++i)
	_M_x[i] = _Private::mod<_IntType, 1, 0, modulus>(lcg());
	
      _M_carry = (_M_x[long_lag - 1] == 0) ? 1 : 0;
      _M_p = 0;
    }

  
  //
  // This implementation differs from the tr1 spec because the tr1 spec refused
  // to make any sense to me:  the exponent of the factor in the spec goes from
  // 1 to (n-1), but it would only make sense to me if it went from 0 to (n-1).
  //
  // This algorithm is still problematic because it can overflow left right and
  // center.
  //
  template<typename _IntType, _IntType __m, int __s, int __r>
    template<class Gen>
    void
    subtract_with_carry<_IntType, __m, __s, __r>::
    seed(Gen& gen, false_type)
    {
      const int n = (std::numeric_limits<_IntType>::digits + 31) / 32;
      for (int i = 0; i < long_lag; ++i)
	{
	  _M_x[i] = 0;
	  unsigned long factor = 1;
	  for (int j = 0; j < n; ++j)
	    {
	      _M_x[i] += gen() * factor;
	      factor *= 0x80000000;
	    }
	  _M_x[i] = _Private::mod<_IntType, 1, 0, modulus>(_M_x[i]);
	}
      _M_carry = (_M_x[long_lag - 1] == 0) ? 1 : 0;
      _M_p = 0;
    }
  

  template<typename _IntType, _IntType __m, int __s, int __r>
    typename subtract_with_carry<_IntType, __m, __s, __r>::result_type
    subtract_with_carry<_IntType, __m, __s, __r>::
    operator()()
    {
      // derive short lag index from current index
      int ps = _M_p - short_lag;
      if (ps < 0) ps += long_lag;
      
      // calculate new x(i) without overflow or division
      _IntType xi;
      if (_M_x[ps] >= _M_x[_M_p] + _M_carry)
	{
	  xi = _M_x[ps] - _M_x[_M_p] - _M_carry;
	  _M_carry = 0;
	}
      else
	{
	  xi = modulus - _M_x[_M_p] - _M_carry + _M_x[ps];
	  _M_carry = 1;
	}
      _M_x[_M_p++] = xi;
      
      // adjust current index to loop around in ring buffer
      if (_M_p >= long_lag)
	_M_p = 0;
      
      return xi;
    }

  
  template<class _E, int __p, int __r>
    typename discard_block<_E, __p, __r>::result_type
    discard_block<_E, __p, __r>::
    operator()()
    {
      if (_M_n >= used_block)
	{
	  while (_M_n < block_size)
	    {
	      _M_b();
	      ++_M_n;
	    }
	  _M_n = 0;
	}
      ++_M_n;
      return _M_b();
    }

_GLIBCXX_END_NAMESPACE
}