int vs long

[Skaperen]

in Python3 i'd like to determine if an int is large enough that its value would have to be long in Python2.

[Larz60+]

but long integers in python 2 are not the same as a double in 'C':

Long integers have unlimited precision. Floating point numbers are usually implemented using double in C

from: https://docs.python.org/2/library/stdtypes.html

[Gribouillis]

Here is what I got in a python 2 console

Python 2.7.12 (default, Dec  4 2017, 14:50:18) 
[GCC 5.4.0 20160609] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import sys
>>> sys.maxint
9223372036854775807
>>> sys.maxint + 1
9223372036854775808L
>>> sys.maxint + 1 == 2 ** 63
True
>>> - sys.maxint - 1
-9223372036854775808
>>> - sys.maxint - 2
-9223372036854775809L

So a good test seems to be

>>> def int_suffices(n):
...     return - 2 ** 63 <= n < 2 ** 63
...
# OR
>>> def int_suffices(n): return n >> 63 in (-1, 0)
...

[Skaperen]

i guess i have to consider that a Pyton3 implementation might do it all different than a Pythin2 implementation. i was looking to know a limit to run at for speed, but this could vary by implementation, all else being the same, about the only thing i can count on is that the limit won't be less than 127.

[Gribouillis]

this could vary by implementation

In python 3, you could perhaps get a more general result with

import sys
BLEN = sys.maxsize.bit_length()
def int_suffices(n):
    return n >> BLEN in (-1, 0)