Adds pow alias for python newcomers.

[gradha]

No description provided.

[reactormonk]

You forgot the actual code.

[gradha]

Added code.

[Araq]

Make it work for integers too. And I prefer an inline proc.

[gradha]

Made inline proc. Added also to complex module. How is this to work with integers? I see no pow for int.

[reactormonk]

math.h only defines it for float.

[Araq]

So implement something like:

proc pow(a, b: int): int =
  var a = a
  var b = b
  result = 1
  while true:
    if (b and 1) == 1:
      result *= a
    b = b shr 1
    if b == 0: break
    a *= a

[reactormonk]

proc pow(a, b: int): int =
  var a = a
  var b = b
  result = 1
  while true:
    if (b and 1) == 1:
      result *= a
    b = b shr 1
    if b == 0: break
    a *= a

for i in 1..10000000:
  var foo = pow(2, 30)

gives me

➜  /tmp  time ./inttest  
./inttest  1.48s user 0.00s system 98% cpu 1.493 total

vs. casting

import math 
for i in 1..10000000:
  var f = 2
  var p = 30
  var foo = int(pow(float(f), float(p)))

which results in

➜  /tmp  time ./floattest
./floattest  0.71s user 0.00s system 98% cpu 0.722 total

[zielmicha]

The float version is faster, but incorrect for numbers over 2**60

[Varriount]

You could always have it use the second method for numbers above 2**60, if you're especially concerned about speed.

[Araq]

The benchmark is flawed, you need to try different exponents and weight them by a reasonable likelihood. Hint: exponent 3 is more likely than exponent 30...

[Araq]

** is weird, it should be ^