ANN: H98 FFI Addendum 1.0, Release Candidate 10

[Malcolm Wallace]

"Simon Marlow" <simonmar at microsoft.com> writes:

> Are you suggesting we ought to be able to define new primitive types?

Sorry, yes indeed that was the suggestion.

Or at least, it is not so much a suggestion, as a recognition that
Hugs and nhc98 already use the 'data T' mechanism for this purpose,
so we should consider if there might be possible confusion when
adding a mechanism to the FFI standard which has the same syntax but
a different purpose.

One design alternative would be to extend the FFI standard to allow
the addition of new primitive types.

Another design alternative would be to use different names for
the primitive datatype mechanism and the empty datatype mechanism.
(Way back when, nhc13 had "data primitive Int" for the former instead
of "data Int", and Alastair has just suggested "foreign import data T"
as a possible syntax for the latter.)

Regards,
    Malcolm