After not making much progress on bpython for a while, someone in #bpython asked for better paste support, and I found I needed the same feature. Building something someone else needs and building something I want are two of my favorite things!
I’d like to give a presentation this week at Hacker School about features of Python folks might not have seen, in the form of short, standalone code snippets we could try to predict the output of together. (part Gary Bernhardt’s Wat, part Coding Bat, part http://rubykoans.com/) I’d like to be quickly jumping back and forth between a REPL and my examples. I considered both IPython notebook and what I think of as an emacs-style REPL (an editor in one vertical split and REPL in the other), but settled on good old bpython, which still just feels right to me for interactive exploration of Python.
I previously added a send-to-editor key to my bpython fork, but to
quickly get my examples into a REPL, I wanted better paste support and the
ability to automatically paste the contents of a file in: now
bpython-curtsies --type example.py. I’ve
got a folder of small example files I can run individually or examine in more
detail in the REPL.
The content is going to be all over the place, with the general pattern of “Do you understand this code? You should, it’s awesome! (see this reference if you don’t)". It will have examples as simple as “functions are first class - look at me storing them in a datastructure,” as well as more subtle things like how hoisting works in Python:
a = 1 def foo(): print a a = 2 foo() #UnboundLocalError: local variable 'a' referenced before assignment
I have tried to cut Python language topics roughly in half, so won’t touch in this session on object orientation, classes, attribute lookup, dunder methods, or modules. If this session is useful to folks I’ll do those in another one covering those.
I aim to write up some notes soon, because otherwise people might take notes when I recommend references, and I’d prefer we all spend the time playing with a REPL :)
Hopefully the session gives people language to use to describe issues in their code, reminds them that better techniques may well exist, and advertises references I’ve found particularly useful like “Python Essential Reference”, Dive into Python, and Loop like a native. When anyone asks me how to learn Python most quickly, my answer is still to write a lot of code and get feedback on it, but I hope a few pinprick glimpses of cool language features will be useful too.