Bug report

Bug description:

The command-line interface of random documents that, when no option is given, an integer input is handled "same as --integer". For an integer that is zero or negative, it prints a random float instead:

$ python -m random -i 0
Traceback (most recent call last):
  ...
ValueError: empty range in randint(1, 0)
$ python -m random 0
0.0
$ python -m random -5
-2.8133945085995364

main() calls randint() inside the same try block that converts the input with int():

try:
    # Is it an integer?
    val = int(val)
    return randint(1, val)
except ValueError:
    try:
        # Is it a float?
        val = float(val)
        return uniform(0, val)

so the ValueError that randint() raises for an empty range is taken to mean that the input is not an integer, and the input is then handled as a float. Keeping only the int() call inside the try block makes python -m random N behave like python -m random --integer N for every integer.

CPython versions tested on:

3.14, 3.15, CPython main branch

Operating systems tested on:

Linux

Linked PRs

• gh-161452: Fix python -m random with a non-positive integer input #161453