[pypy-dev] curses_cffi and others

[Armin Rigo]

Hi Gabriel,

On Fri, Mar 1, 2013 at 4:22 PM, Gabriel de Perthuis <g2p.code at gmail.com> wrote:
> Packages using CFFI should be able to install-require CFFI whatever the
> interpreter is, including requiring newer versions than the one the interpreter
> ships.

The problem is that the pure Python CFFI is rather tied to the
extension module _cffi_backend (e.g. you can't add new methods to
cdata objects from the pure Python side).

So the problem is really: it's not possible to install a newer CFFI on
top of an existing PyPy interpreter at all (unless you install a newer
_cffi_backend.so as a cpyext C extension module, which completely
defeats the performance point).  But the idea is that CFFI should soon
be more or less frozen; I don't want to keep growing its feature set
forever.

Even so, more generally, I think that requiring users to install a
newer version of the interpreter is actually feasible, when we are
talking about PyPy and not CPython.  A "newer version" of PyPy is just
an internal improvement, but not a language change --- it's always
2.7.  I think that so far, our users are (or should be) ready to deal
with this minor annoyance; if they are not and really want a
completely stable, never-changing interpreter, then they're going to
stick with CPython anyway.

In other words, I regard this as similar to: "this program requires
PyPy 2.0 and not a lower version because otherwise its usage of
greenlets is going to be a huge performance hit".


A bientôt,

Armin.