I learned some neat stuff at clepy last night

Brian Beck showed how to use metaclasses and descriptors to make DSLs with python.

I do this kind of this kind of thing every so often in my code:

def f(x):
class C(object):
y = x
return C

That function takes a parameter and makes and returns a class based on that parameter. Whoop-di-do. I was surprised to learn that you can’t do this:

class C(object):
x = 99
class D(object):
y = x + 1

I gotta explore this some more until it makes sense.

Here’s another neat trick: It isn’t possible to add two classes together:

>>> class C(object):
... pass
...
>>> C + C
------------------------------------------------------------
Traceback (most recent call last):
File "", line 1, in
TypeError: unsupported operand type(s) for +: 'type' and 'type'

But if you want to support this, the solution would be to define an __add__ method on the metaclass:

>>> type(C)

>>> class MC(type):
... def __add__(self, other):
... print 'Adding!'
... return 99
...
>>> class C(object):
... __metaclass__ = MC
...
>>> C + C
Adding!
99

Wacky, right? More realistically, I could build a new class by taking attributes of both classes together. In other words, if class C has a class attribute x, and class D has a class attribute y, then we can use a metaclass to add C and D together to get a new class E, that has both x and y as class attributes.

In this example, C has a class attribute x and D has a class attribute y. When I add the two classes, I get a new class with both of those class attributes.

>>> C.x, D.y
(99, 98)
>>> E = C + D
>>> E.x, E.y
(99, 98)

Here’s the metaclass that allows this sort of nonsense:

class MC(type):

def __add__(self, other):

class E(self):
pass

for k,v in other.__dict__.items():
if k not in ('__dict__', ):
setattr(E, k, v)

return E

Google presentation at Clepy on August 6th, 2007

Tonight Brian Fitzpatrick (Fitz) from the Chicago Google office did a presentation for the clepy group on version control at Google. They use subversion on top of their own super-cool bigtable filesystem back end.

We had a good discussion on the merits of centralized vs. decentralized version control. According to Fitz, decentralized systems discourage collaboration. He made the joke, “Did you hear about the decentralized version control conference? Nobody showed up.” He made the point that centralized repositories encourage review and discussion. I agree with that.

Apparently subversion 1.5, which will be released in a few months, will have much improved merging facilities. We won’t need to use --stop-on-copy to figure out where we branched. Also, it will be safe to repeat a merge, because nothing will happen on the second attempt.