[Fwd: cvs commit: fptools/libraries/base/Foreign/C Types.hs fptools/libraries/base/System CPUTime.hsc Directory.hs Time.hsc fptools/libraries/base/System/Posix Types.hs fptools/libraries/base/include CTypes.h]

[Sven Panne]

I wrote:
> I've just commited a change to the fptools repository, proposing a minor 
> change in the FFI spec, see below. Comments welcome. [...]

My previous mail was a little bit terse, so I'd like to explain my proposal
in more detail: A part of the FFI spec describes a mapping of some C types to
Haskell types (CInt & friends), including the instances defined for them. The
relevant C types are (in C99-speak) "integer types" (char, short int, ...),
"floating types" (float, double, long double), and "arithmetic types" (= union
of integer and floating types). The Haskell instances for the corresponding
types should mirror what one can do with the C types. This is almost the case,
but my point is that for the arithmetic types CTime and CClock (= clock_t/time_t),
Real instances are missing in the spec. Having only Num available, one can
convert an Integer (and other Integral types via fromIntegral) to such a type,
but not the other way round, although this is perfectly possible in C. Having
a Real instance available, too, one could do this, e.g. via:

    round . fromRational . toRational

This would mirror being able to say in C:

    time_t t = ...
    int i = (int)t;

In a nutshell: I propose to add Real instances to CTime and CClock. Additionally,
the wording of the relevant part of the FFI spec should probably use the same
terms as C99 (integer/floating/arithmetic types), define the instances for these
kind of types, and say e.g. that CTime is an arithmetic type. This would make
things a bit clearer and less redundant.

Cheers,
    S.