[cyber_kinetist]

Why should I do this, when I can just write:

    for (int i = 0; i < nums.size(); i++) {
        nums[i] = rng();
    }

Simple, easy to understand, actually compiles to efficient code even in Debug mode, don't need to include <algorithm> header which includes 10000+ lines of STL code which bloats compile times, etc.

[tsimionescu]

Even better, so you can't accidentally move out of bounds:

  for (int i : nums) {
      nums[i] = rng();
  }

Edit: Ooops, I'm an idiot, this doesn't do anything useful (i is every value in nums, not the indices of nums...)

[leshow]

Doesn't that not work because `nums` is all initialized to 0 though?

Like, wouldn't it be

   for (int num : nums) {
      num = rng();
   }

(I don't know any c++). I kind of expect the above not to work though because it's doing an assignment. I suppose this is a good argument to just use a `fill` or `generate` function...

[tsimionescu]

You're right, I completely forgot what it meant... The loop I wrote was basically equivalent to :

  for (int i = 0; i < nums.size(); ++i) {
    nums[nums[i]] = rng();
  }

Which happens not to go out of bounds simply because nums[i] is initially 0 for every i in the range - so it sets nums[0] to a different random number 1000 times.