[docs] Context managers, file.writelines, etc

[delirious-lettuce]

• Use file.writelines instead of file.write for each line
• Use with statement context managers instead of manually closing each file
• Use a try/except inside main to unpack the three sys.argv arguments (raise ValueError if not enough/too many arguments)
  • Except for gosecure_client_install where only one argument is needed

---

I noticed quite a few blocks of code like this and thought it might be inefficient.

with open("/etc/modules", "r+") as f:
    for line in f:
        if "bcm2708-rng" in line:
            break
    else: # not found, we are at the eof
        call("sudo sh -c 'echo bcm2708-rng >> /etc/modules'")

It seemed odd to iterate through each separate line of f and check for the exact same string. I set up a small test to get an idea of how long it took compared to just searching f.read() for the one string.

In[20]: with open('treasure_test_output.txt', 'w') as f:
   ...:     f.writelines([('*' * 150) + '\n' for _ in xrange(999)])
   ...:     f.write(('*' * 150) + 'treasure\n')
   ...: 
   ...: one = """\
   ...: with open('treasure_test_output.txt', 'r') as f:
   ...:     for line in f:
   ...:         if 'treasure' in line:
   ...:             break
   ...: """
   ...: 
   ...: two = """\
   ...: with open('treasure_test_output.txt', 'r') as f:
   ...:     'treasure' in f.read()
   ...: """
In[21]: timeit(one, number=100000)
Out[21]: 
36.9743390083313
In[22]: timeit(two, number=100000)
Out[22]: 
10.249281167984009

To test the worst case scenario, I created a 1000 line file where the word "treasure" was at the end of the very last line. Using content = f.read() and searching for that one string was around 3 times faster every time I checked it.

I went through and changed the format of each of these instances to search for the desired string in the f.read() string.