On 23 Jan 2014 18:32, "Suresh V." <suresh_vv@yahoo.com> wrote:
>
> Nicely done :-)
>
> "foo" may come from a library or something, so rather than a decorator we may have to monkey patch it. Unless there is a nicer solution.
>
> Will functools be a good place for something like this?

Another idea along similar lines is the object model in Elk: http://frasertweedale.github.io/elk/ (that's a before/after/around subclass method model, designed specifically as an alternative to using super() to call up to the parent implementation).

The main problem with the idea of doing this as a more general feature for arbitrary callables is that it has most of the same downsides as monkey-patching while being strictly less powerful and even more confusing (since it would be difficult to model clearly in tracebacks).

Cheers,
Nick.

>
>
> On Thursday 23 January 2014 01:50 PM, Chris Angelico wrote:
>>
>> On Thu, Jan 23, 2014 at 7:11 PM, Suresh V. <suresh_vv@yahoo.com> wrote:
>>>
>>> On Thursday 23 January 2014 01:22 PM, Chris Angelico wrote:
>>>>
>>>>
>>>> On Thu, Jan 23, 2014 at 6:20 PM, Suresh V. <suresh_vv@yahoo.com> wrote:
>>>>>
>>>>>
>>>>> Can we add these two attributes for every function/method where each is a
>>>>> list of callables with the same arguments as the function/method itself?
>>>>>
>>>>> Pardon me if this has been discussed before. Pointers to past discussions
>>>>> (if any) appreciated.
>>>>
>>>>
>>>>
>>>> I'm not exactly sure what you're looking for here. What causes a
>>>> callable to be added to a function's __before__ list, and/or what will
>>>> be done with it?
>>>
>>>
>>>
>>> These are modifiable attributes, so something can be added/deleted from the
>>> __before__ or __after__ lists.
>>>
>>>
>>>>
>>>> If you mean that they'll be called before and after the function
>>>> itself, that can be more cleanly done with a decorator.
>>>
>>>
>>>
>>> Yes. Each item in the list will be called in order immediately before/after
>>> each invocation of the function. This is kinda like decorators, but more
>>> flexible and simpler. Scope for abuse may be higher too :-)
>>
>>
>> def prepostcall(func):
>>      def wrapper(*args,**kwargs):
>>          for f in wrapper.before: f(*args,**kwargs)
>>          ret = func(*args,**kwargs)
>>          for f in wrapper.after: f(*args,**kwargs)
>>          return ret
>>      wrapper.before = []
>>      wrapper.after = []
>>      return wrapper
>>
>> @prepostcall
>> def foo(x,y,z):
>>      return x*y+z
>>
>> foo.before.append(lambda x,y,z: print("Pre-call"))
>> foo.after.append(lambda x,y,z: print("Post-call"))
>>
>> Now just deal with the question of whether the after functions should
>> be called if the wrapped function throws :)
>>
>
>
>
>
>> ChrisA
>> _______________________________________________
>> Python-ideas mailing list
>> Python-ideas@python.org
>> https://mail.python.org/mailman/listinfo/python-ideas
>> Code of Conduct: http://python.org/psf/codeofconduct/
>>
>
>
> _______________________________________________
> Python-ideas mailing list
> Python-ideas@python.org
> https://mail.python.org/mailman/listinfo/python-ideas
> Code of Conduct: http://python.org/psf/codeofconduct/