Re: [ctypes-users] Re: Future plans for ctypes

[Lenard L. <le...@te...>]

On 22 Feb 2006 at 23:09, Paul Moore wrote:

> On Wed, 22 Feb 2006 19:36:36 +0100, Thomas Heller <th...@py...>
> wrote:
> 
> >Lenard Lindstrom wrote:
> >> Congratulations on getting ctypes into the Python distribution.
> 
> Agreed.
> 
> >> I have been following the discussed changes and this is just an idea. 
> >> Why not release ctypes into the Python core library under a new name. 
> >> There is certainly a precident with optparse.  Then it can have the 
> >> interface you want. For backwards compatibility a ctypes wrapper 
> >> package can be built around the core package, like wxPython is now 
> >> built around wx. To me ctypes implies an alternative to the struct 
> >> module. It certainly is not that. Is there some licensing reason it 
> >> can't be named ffi?
> >
> >There's no principal problem, afaik, to rename it to ffi.  But, I don't like
> >this idea, although I will have to sleep over it.  IIRC, optik was renamed
> >from optparse because the python-dev crowd did not like the name.
> 
> I'm not keen on the name ffi. I'm not at all sure why, in many ways
> it's much more accurate. Just gut feel, and the fact that I'm used to,
> and like, the name ctypes.
> 
> The case of optparse/optik was a little different, IMHO. I think optik
> was considered a little too "cute" for a stdlib module name, and
> optparse was more "formal" (my recollection could be wrong, though). I
> don't think that applies here.
> 
optparse was just an example to show that backwards name 
compatibility is not an issue when adding a package to the core 
library. The suggestion was to rename the ctypes release version so 
that a "ctypes" wrapper module, with all the deprecated features, 
could continue to exist for backwards compatibility. "ffi" was just a 
suggestion; any name would do. But  "ctypes" is fine with me.

Lenard Lindstrom
<le...@te...>