[Python-Dev] Should vars() return modifiable dict?

[Terry Reedy]

On 10/3/2012 11:34 AM, Steven D'Aprano wrote:
> On 03/10/12 18:54, Serhiy Storchaka wrote:

>> Should behavior of vars() be corrected for locals?
>
> I believe you are misinterpreting what you are seeing. In this case,
> vars() simply returns locals(), which is an ordinary dict, but changes
> to that dict are not guaranteed to propagate to the actual local
> variables themselves. You make changes to that dict, then call vars()
> again, which returns a fresh locals() dict. So what you are seeing is
> simply a side-effect of the fact that changes to locals() may not
> actually effect the local variables.
>
> Note that in IronPython, the behaviour of your code is different:
>
>
> steve at runes:~$ ipy
> IronPython 2.6 Beta 2 DEBUG (2.6.0.20) on .NET 2.0.50727.1433
> Type "help", "copyright", "credits" or "license" for more information.
>>>> def f():
> ...     x = 42
> ...     print(vars())
> ...     vars()['x'] = 43
> ...     vars()['y'] = 44
> ...     print(x, vars())
> ...
>>>> f()
> {'x': 42}
> (43, {'y': 44, 'x': 43})
>
>
>
>> Should vars() for objects with __slots__ [1] returns modifiable or
>> non-modifiable dict?
>>
>> [1] http://bugs.python.org/issue13290
>
>
> You are assuming that the behaviour of vars(obj) should change. I don't
> think
> that is necessarily the case.
>
> vars(obj) is defined as returning the object __dict__ attribute. If an
> object
> has no __dict__, vars() should (pick one):
>
> 1) Keep the current behaviour and raise an exception.
>
> 2) Return a regular dict containing {slot: value} for each of the slots.
>     Since the dict is a copy, changes to the dict will not effect the
>     original object.
>
> 3) Return a dictproxy containing {slot: value} for each of the slots.
>     Since the dictproxy does not support item assignment, you can't
>     modify it.
>
> 4) Return some other proxy object such that changes to the dict will
>     also change the object's slot attributes.
>
>
> I find myself unable to choose between 2) and 4), which suggests that
> the status quo wins and we keep the current behaviour.
>
>
>


-- 
Terry Jan Reedy