Re: ffi library port to Windows

Enrico Migliore wrote:

> Hi Etienne,
>
>> Hi Enrico,
>>
>> Have you given a look at :
>> http://ctypes.cvs.sourceforge.net/ctypes/ctypes/source/libffi_msvc/
>>
>> Apparently, ctypes uses an old version of libffi which they ported to
>> msvc.  Maybe looking at it could help you.
>>
>> See below.
>>
>>  
>>
> yes, I took a look at that URI and realized that the MSVC port has 
> already been done!
>
> The GCC assembly file win32.s, was translated for the MSVC assembler, 
> in 2002 by John Beniton and Ranjit Mathew.
>
> Tomorrow I'll download the ffi port and try to run ALL the tests. If 
> all tests succeed, I'll start to tackle the task of porting SableVM to 
> MSVC.
>
> Enrico
>
hi Etienne,

 I confirm that a WIN32 port of the ffi library alredy exists. The port 
is part of the ctypes project:

 http://starship.python.net/crew/theller/ctypes/


Before going on, I got 2 questions:


1. directory tree
----------------

The ctypes guys, instead of adding #ifdef's to the original source files 
of ffi, decided to make a copy of the following files:

prep_cif.c
ffi.c
ffi.h
fficonfig.h
ffitarget.h

into new directory, called: libffi_msvc, and then worked on those files.

That means that GCC and MSVC work on different source files. A while 
ago, we agreed on adding #ifdef wherever possible, in order to have just 
one source tree. What are we supposed to do?


2. license issue
---------------
Am I supposed to ask for permission to the Ctpye guys, before re-working 
and submitting the source files to the SableVM project?

Enrico