python vs bash

imiell@basquiat $  cat a.sh 
#!/bin/bash
for i in {1..100000}
do
        VAR='HEADERMy voice is my passwordFOOTER'
        PASS=${VAR#HEADER}
        PASS=${PASS%FOOTER}
done

imiell@basquiat $  cat a.py 
for i in range(100000):
        var = 'HEADERMy voice is my passwordFOOTER'
        password = var.lstrip('HEADER')
        password = password.rstrip('FOOTER')


imiell@basquiat $ time ./a.sh

real	0m2.442s
user	0m2.395s
sys	0m0.026s


imiell@basquiat $ time python a.py 

real	0m0.104s
user	0m0.085s
sys	0m0.010s




imiell@basquiat $  cat a.sh 
#!/bin/bash
for i in {1..1000}
do
        VAR='HEADERMy voice is my passwordFOOTER'
        PASS=${VAR#HEADER}
        PASS=${PASS%FOOTER}
done

imiell@basquiat $  cat a.py 
for i in range(1000):
        var = 'HEADERMy voice is my passwordFOOTER'
        password = var.lstrip('HEADER')
        password = password.rstrip('FOOTER')

imiell@basquiat $  time ./a.sh

real    0m0.035s
user    0m0.034s
sys     0m0.001s

imiell@basquiat $  time python a.py 

real    0m0.029s
user    0m0.016s
sys     0m0.013s

Note that (l|r|)strip() doesn't strip the verbatim strings you pass in, but any number of the characters passed in! That is, .strip('HEADER') is equivalent to .strip('ADEHR'), for instance, and if the password we want to get at happened to start with any of those characters (or ended with any number of E, F, O, T, or R), those would have gotten stripped as well and we'd ended up with the wrong password.

A more equivalent version would be

password = (password[6:] if password.startswith('HEADER') else password)
password = (password[:-6] if password.endswith('FOOTER') else password)

(or you could use the re module).

Fair point - out of context this makes less sense. I was looking for a speed comparison to prove bash is slower. But it wasn't.

Except that the startup time of python makes a difference here - have updated the gist to reflect the fact that bash is slower after all.