Epsilon: random()-random() gets simplified to 0

[lastfuture]

When trying to simulate throwing two dice I stumbled across this:
randint(1,6)+randint(1,6) gets simplified to 2*randint(1,6)

That does not yield the expected result of two individual dice being thrown, but instead of one dice being thrown and counted twice, which is a huge difference.

This also happens in other places, e.g. in Sequences where one might want to visualize a distribution of values when throwing a certain number of dice.

I propose that random integers should be treated as always different so they won't be combined when simplifying the term

[lastfuture]

This also extends to something like random()-random() which always gets simplified to 0

[Ecco]

That's a good catch 😄

[Ecco]

like random()-random() which always gets simplified to 0

Are you sure? Maybe you were just (un)lucky? 😄

[notandxor]

“The generation of random numbers is too important to be left to chance.”
--forgot who

[tarapitha]

Robert R. Coveyou: http://www.quotationspage.com/quote/461.html

[lastfuture]

like random()-random() which always gets simplified to 0

Are you sure? Maybe you were just (un)lucky? 😄

That reminds me that getRandomNumber() would be a funny easter egg
https://xkcd.com/221/

[Ecco]

This has been fixed in 210ec62 😄