> So a cdef int s a py_int, but a cdef float is not a py_float?

No.  Clarifying my above idea -- to the programmer, a cdef int would
be a c int, a cdef float would be a c float, a py_int would be a
python int, and a py_float would be a python float.  In the generated
c code, a cdef py_int could be a c long or c long long with python
operations and a py_float would be a c double.

To step back -- if the only way to drop into fast compiled code is to
learn c semantics, it will cause problems for people who just want to
cythonize their python code.  If you use python semantics, it will
cause problems for people who are c programmers using cython to write
mixed python-c code.  So doing this seems to give everyone something
they like.

I'm also thinking of proposed steps in the cython development such as
automatically determining types (e.g. of a in 'a = 5').  Having a
python integer would make that easier.

--Hoyt

++++++++++++++++++++++++++++++++++++++++++++++++
+ Hoyt Koepke
+ University of Washington Department of Statistics
+ http://www.stat.washington.edu/~hoytak/
+ [email protected]
++++++++++++++++++++++++++++++++++++++++++
_______________________________________________
Cython-dev mailing list
[email protected]
http://codespeak.net/mailman/listinfo/cython-dev

Reply via email to