[Python-Dev] C99 code in the Python core?

["Martin v. Löwis"]

> (1) When is it okay to use C99 code in the Python core?  More particularly,
> is it considered acceptable to use widely-implemented library functions that
> are specified in C99 but not ANSI C, or widely-implemented features that
> are new to C99?

[C99 is also ANSI C, IIUC. ANSI has adopted ISO/IEC 9899:1999 as a U.S.
 national standard.]

It's ok to use functions of the C99 standard library if you have a
configure test and a fall-back implementation, or if you know that the
function is available on all systems we care about.

> Or is C99 code now acceptable pretty much anywhere?

No. As others have pointed out, Microsoft still hasn't implemented in
Visual C.

> If so, should
> PEP 7 be updated?  It currently says: """Use ANSI/ISO standard C
> (the 1989 version of the standard)."""

No.

> (2) Is it okay to use the '%a' format specifier for sprintf, sscanf and friends.
> Are there major platforms where this isn't implemented?  (Using
> '%a' would make the issue implementation much simpler.)

It's implemented in VS 2008, see

http://msdn.microsoft.com/en-us/library/hf4y5e3w.aspx

On the other hand, people still might try to run Python on older
versions of Solaris, such as Solaris 2.6 (which was released 1997).
I don't know when Solaris' CRT first started to support this.

I'd add a configure test, and, at run-time, raise an exception
if the C library doesn't support it yet somebody tries to use it.

Regards,
Martin