Message189175
I've come across something in the implementation here that I'd like some clarification on. What is the purpose of overriding __dir__ in Enum and EnumMeta? It doesn't change any behavior that I'm aware of, just makes things look a little nicer when someone calls dir() on their Enum. And, in fact, it can make things a little confusing. For example:
>>> class Test(enum.Enum):
... foo = 1
... bar = 2
... baz = 3
...
>>> dir(Test)
['__class__', '__doc__', '__members__', 'bar', 'baz', 'foo']
>>> Test.mro
<built-in method mro of EnumMeta object at 0x01D94D20>
This brings up another interesting case:
>>> class Test2(enum.Enum):
... mro = 1
... _create = 2
...
>>> dir(Test2)
['__class__', '__doc__', '__members__', '_create', 'mro']
>>> Test2.__members__
mappingproxy(OrderedDict([('mro', <Test2.mro: 1>), ('_create', <Test2._create: 2>)]))
>>> Test2['mro']
<Test2.mro: 1>
>>> Test2.mro
<built-in method mro of EnumMeta object at 0x01D90210>
>>> Test2._create
<bound method type._create of <class 'enum.EnumMeta'>>
>>>
From using "mro" or "_create", I would have expected either ValueError or for them to work properly. I don't know whether this should be fixed (one way or the other), documented, or just left alone; those kind of names really shouldn't ever be used anyway. It's something I stumbled across, though, and I just wanted to make sure that those who do have opinions that matter are aware of it :) |
|
Date |
User |
Action |
Args |
2013-05-13 20:17:26 | zach.ware | set | recipients:
+ zach.ware, gvanrossum, barry, ncoghlan, benjamin.peterson, ezio.melotti, alex, eli.bendersky, docs@python, ethan.furman, python-dev, pconnell, isoschiz |
2013-05-13 20:17:26 | zach.ware | set | messageid: <1368476246.75.0.836362831038.issue17947@psf.upfronthosting.co.za> |
2013-05-13 20:17:26 | zach.ware | link | issue17947 messages |
2013-05-13 20:17:26 | zach.ware | create | |
|