result type of calling an extension type is unknown

[scoder]

For an extension type MyExt, calling MyExt() can only result in an instance of type MyExt. However, Cython does not assign the type to the result and uses the generic Python object type instead, so that things like MyExt().some_c_func() fail.

[robertwb]

Are we saying we won't ever provide new support?

[scoder]

The discussion on cython-dev showed that there are valid use cases for returning objects of a different type from a future __new__ special method. Therefore, this ticket turns into a request for optimising the most common case that calling a known type returns an instance of exactly the called type or at least a subtype.

[cwitty]

I'm copying this here from cython-dev:

How about this? In the hypothetical future version of Cython that allows overriding __new__, say that the overrides have to be declared in the .pxd file, with

   cdef TYPE __new__(self, foo, bar)

(where TYPE defaults to the class being declared, if omitted). So when generating code, if the class doesn't override __new__, you know that calling the class returns an instance; if the class does override __new__, then the return type is explicitly given. This could be a parent class, for my use case where I want to return an instance of a sibling class; or it could be "object", if anybody ever has a use case for having __new__ return an arbitrary object.

This idea means we can go ahead and optimize the common case as initially requested, without worrying about compatibility headaches with future Cython versions.

[scoder]

The thread leading up to this is here:

 http://comments.gmane.org/gmane.comp.python.cython.devel/4849?set_lines=100000

[scoder]

Supporting __new__() in extension types is ticket #238.

[scoder]

Cython now handles extension type tracking like Pyrex. See ticket #238 for ___new__() support.