!standard  B.1  (49)                                   03-01-02  AI95-00320/01
!class binding interpretation 03-01-02
!status work item 03-01-02
!status received 03-01-02
!qualifier Omission
!priority Low
!difficulty Easy
!subject Violating Ada semantics with an interfacing pragma

!summary

The note B.1(49) ("an interfacing pragma might result in an effect
that violates Ada semantics.") is correct.

!question

The minutes of the October 2002 ARG meeting say:

  If you violate any Ada semantic requirements in C (by stomping on memory, for
  instance), all bets are off.

This is clearly a requirement, as the Ada compiler has no control over what the
foreign language code can do.

However, the only mention of this in the Standard is a note, B.1(49):

  An interfacing pragma might result in an effect that violates Ada semantics.

It appears that the standard needs a normative statement to this effect.

!recommendation

(See Wording.)

!wording

Add after B.1(38):

A convention other than the ones defined by the language in subclause 6.3.1
are known as a foreign convention. The effect of a partition containing an
interfacing pragma with a foreign convention may differ from that defined by
this International Standard.

!discussion

!corrigendum B.1(38)

@dinsa
Notwithstanding what this International Standard says elsewhere, the
elaboration of a declaration denoted by the @fa<local_name> of a @fa<pragma>
Import does not create the entity. Such an elaboration has no other effect than
to allow the defining name to denote the external entity.
@dinst
A convention other than the ones defined by the language in subclause 6.3.1
are known as a @i<foreign convention>. The effect of a partition containing an
interfacing pragma with a foreign convention may differ from that defined by
this International Standard.

!ACATS test

We can't usefully test if a compiler does this, as it is a permission, not a
requirement.

!appendix

****************************************************************