Comments on: Python Quirks in Cmd, urllib2, and decorators

By: srid

srid — Sat, 07 Nov 2009 14:42:42 +0000

cmdln is a Python library that does what Cmd and optparser combined .. misses. http://code.google.com/p/cmdln

By: m0j0

m0j0 — Thu, 29 Oct 2009 12:34:37 +0000

Thanks for this — I haven’t actually implemented this in my code yet, so nice to know in advance, though I did expect that to some degree I was still going to have to restrict the user in some way to make parsing the args line predictable.

Another option I forgot to mention that either a) alleviates the issue or b) trades one problem for another is to use raw_input, which Cmd will handle in a sane way. That’s actually what I’m using for my ‘connect’ method now. I imported getpass to handle the password

By: j_king

j_king — Thu, 29 Oct 2009 12:34:21 +0000

Roger is right, you should be using the shlex module. The unicode issue is kind of annoying but can be abstracted away.

Decorators are really simple but I think the trip up is that people expect them to be complex. Just gotta remember that they’re only syntactic sugar. Instead of always calling f(x()) you can just use the decorator syntax. They’re the same thing.

You can get more complex than that, but the beginning Python programmer doesn’t need to worry about it; they’re pretty far edge cases that a programmer should have enough experience with Python to not be phased by it when they come across it.

By: Roger

Roger — Thu, 29 Oct 2009 04:28:44 +0000

As you point out the interesting part is splitting the string into pieces. Using regular split is a bad idea since it won’t cope with quoted arguments.

Instead you can use shlex.split which does the job really well. It has two gotchas. The first is that it doesn’t process backslashes. so if someone did –password=abc\ def then you’ll still get that literally – ie you’ll need to turn backslash space in to a space, t into tab, quote into a quote etc.

The second problem is that shlex.split is completely broken in Python 2 (Py 3 is ok) if you supply a Unicode string. It works on the raw Unicode binary bytes giving a resulting mess. If you are using unicode (which you should) then you have to convert to utf8 first, call it and then convert the results back to Unicode.