------- Comment #6 from ludovic at ludovic-brenta dot org 2007-12-04 20:35
-------
The fact that Boolean'Base is unconstrained _doesn't_ imply
that we don't have a discrete_subtype_indication. Discreteness
and constrainedness aren't the same thing, and one doesn't
imply the other. Boolean'Base is discrete because all Boolean
types are discrete. A discrete_subtype_indication is a
subtype_indication that happens to be discrete, and a subtype_indication
is defined in RM 3.2.2(3) as a subtype_mark with an optional
constraint, and a subtype_mark is defined in 3.2.2(4) as
a subtype_name, which is a name that denotes a subtype, and
a name is defined in RM 4.1(2) as including an attribute_reference.
So this is still a bug.
--
ludovic at ludovic-brenta dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|WAITING |NEW
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15799
