Fix normal_random behavior on negative parameters

[ranisalt]

Pull Request Prelude

• I have followed proper The Forgotten Server code styling.
• I have read and understood the contribution guidelines before making this PR.
• I am aware that this PR may be closed if the above-mentioned criteria are not fulfilled.

Changes Proposed

When the algorithm was updated to avoid skew #4539 I accidentally broke passing negative numbers in reverse order (e.g. calling normal_random(-1, -100) instead of normal_random(-100, -1)) due to the way std::minmax works: it returns references to the minimum and maximum element, and combined with std::tie it would replace minNumber in-place before it's value was moved to maxNumber.

Storing the values in local variables fixes the problem without adding an explicit branch + swap as proposed in the comments.

Issues addressed: fixes #4554