modifying min/max for param_type in a distribution

[bluefrog]

How can I modify the minimum and maximum for the range, that was used during construction of a uniform distribution?

I get a compile error with the following:

#include <iostream>
#include <random>

int round_to_hundred(int d) {
	// return ((d % 100) < 50) ? d - (d % 100) : d + (100 - (d % 100));
  return (d + 50) / 100 * 100;
}

int main() {
  
  int start{300}, end{2000};
  std::random_device rd; // obtain a random number from hardware
  std::mt19937 eng(rd()); // seed the generator
  
  std::uniform_int_distribution<>::param_type pt(start, end);
  std::uniform_int_distribution<> distr(pt); // define the range

  std::cout << "start " << distr.a() << " end " << distr.b() << "\n";
  std::cout << "min " << distr.min() << " max " << distr.max() << "\n";
  std::cout << "Sample Values: " << distr(eng) << ',' << distr(eng) << ',' << distr(eng) << ',' << distr(eng) << std::endl;

  int n {distr(eng)};
	std::cout << n << " " << ((n + 50) / 100 * 100)   << "\n"; // round to nearest hundred

  start = 1; end = 100;
  pt.min(start);
  pt.max(end);
  
  std::cout << "Sample Values: " << distr(eng) << ',' << distr(eng) << ',' << distr(eng) << ',' << distr(eng) << std::endl;
  n = distr(eng);
	std::cout << n << " " << ((n + 5) / 10 * 10)   << "\n"; // round to nearest ten


  return 0;
}

[bluefrog]

Partially solved by constructing a new parameter type, with the new range.
In addition, since the default parameter type still applies to the distribution, one has to override the default with the new param_type by explicitly passing it, as in line 21 & 22.

So this is a working version, but not ideal:

#include <iostream>
#include <random>

int main() {
  
  int start{300}, end{2000};
  std::random_device rd; // obtain a random number from hardware
  std::mt19937 eng(rd()); // seed the generator
  
  std::uniform_int_distribution<>::param_type pt(start, end);
  std::uniform_int_distribution<> distr(pt); // define the range

  std::cout << "start " << distr.min() << " end " << distr.max() << "\n";
  std::cout << "Sample Values: " << distr(eng) << ',' << distr(eng) << "\n";

  int n {distr(eng)};
  std::cout << n << " rounded to nearest 100 " << ((n + 50) / 100 * 100)   << "\n"; // round to nearest hundred

  start = 1; end = 100;
  std::uniform_int_distribution<>::param_type pt2(start, end);
  std::cout << "Sample Values: " << distr(eng, pt2) << ',' << distr(eng, pt2) << "\n";
  n = distr(eng, pt2);
  std::cout << n << " rounded to nearest 10 " << ((n + 5) / 10 * 10)   << "\n"; // round to nearest ten

  return 0;
}

Can anybody suggest a better way ?